{
  "name": "GratisAI Index dataset",
  "description": "Free and open-source AI tools with the pricing label we verified, a one-line reason for that label, and the date it was checked against the vendor page.",
  "source": "https://gratisai.com.au",
  "license": "CC BY 4.0",
  "license_url": "https://creativecommons.org/licenses/by/4.0/",
  "attribution": "GratisAI Index (gratisai.com.au)",
  "generated_at": "2026-11-18",
  "count": 66,
  "fields": {
    "pricing_model": "One of: 100% Free, Generous Free Tier, Limited Free Tier, Open Source, Free Trial. Definitions: https://gratisai.com.au/methodology",
    "pricing_reasoning": "Why the tool carries that label, with the published limits where they exist.",
    "last_verified": "ISO date the entry was last checked against the vendor pricing or docs page.",
    "sources": "Vendor pages the label was based on.",
    "scope_caveat": "Present when the tool is adjacent to the directory scope; explains what it is not."
  },
  "tools": [
    {
      "id": "freebuff",
      "name": "Freebuff",
      "tagline": "Open-source multi-agent coding agent & app builder with CLI, desktop, and cloud sandboxes",
      "description": "Freebuff is an open-source, multi-agent AI coding assistant and full-stack app builder positioned as a free alternative to Claude Code, Cursor, and Devin. No API key or credit card is required: advertising funds a daily allowance of usage hours, and optional paid plans add hours and models. Specialized sub-agents handle context gathering, planning, editing, and automated testing across CLI, Desktop, and Cloud environments.",
      "website": "https://freebuff.com",
      "category": "appbuilder",
      "subcategories": [
        "Multi-Agent",
        "App Builder",
        "CLI",
        "Desktop",
        "Cloud Sandbox"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free plan costs $0 with no card and refills 30 hours of model usage daily with unlimited messages per session; paid plans start at $8/month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://freebuff.com/plans",
        "https://freebuff.com/blog/how-is-freebuff-free"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "CLI",
        "MacOS",
        "Windows",
        "Linux",
        "Web"
      ],
      "pros": [
        "No credit card or API key required; ads fund the free daily hours",
        "Multi-agent architecture divides tasks between planning, coding, and review",
        "Open-source CLI, agents, and SDK under Apache 2.0 on GitHub"
      ],
      "cons": [
        "Free usage is metered in daily hours that vary by country"
      ],
      "date_added": "2026-10-15",
      "scope_caveat": null
    },
    {
      "id": "bolt-new",
      "name": "Bolt.new by StackBlitz",
      "tagline": "Prompt, build, run, and deploy full-stack web applications directly in the browser",
      "description": "Bolt.new leverages WebContainers to run a full Node.js environment directly in your browser tab. Prompt the AI to build complete full-stack React, Next.js, Vue, or Svelte applications with live package installation, hot-reload terminal, and instant deployment.",
      "website": "https://bolt.new",
      "category": "appbuilder",
      "subcategories": [
        "WebContainer",
        "Full-Stack",
        "Browser IDE",
        "Live Preview"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free forever, no card: 1M tokens/month with a 300K daily cap, 3 collaborators and hosting; Pro is $25/month for 10M tokens and a custom domain.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://bolt.new/pricing"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Runs complete Node.js environments and npm packages in-browser",
        "Live interactive preview updates in real-time as the agent codes",
        "Generous free daily token allocation for building full apps"
      ],
      "cons": [
        "Requires internet access to stream model tokens"
      ],
      "date_added": "2026-09-05",
      "scope_caveat": null
    },
    {
      "id": "openhands",
      "name": "OpenHands (All-Hands AI)",
      "tagline": "Open-source autonomous AI software development agent capable of complex engineering",
      "description": "OpenHands (formerly OpenDevin) is the flagship open-source autonomous software engineering platform. It can write code, run terminal commands, fix bugs, read pull requests, and browse the web inside secure Docker environments.",
      "website": "https://www.openhands.dev",
      "category": "appbuilder",
      "subcategories": [
        "Autonomous Agent",
        "Docker",
        "Open Source",
        "Devin Alternative"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed agent you self-host with your own LLM key, so there is no vendor allowance and no card; cloud offers a free $10 credit then Teams at $500/month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.openhands.dev/pricing",
        "https://github.com/All-Hands-AI/OpenHands"
      ],
      "license": "MIT",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows",
        "Web"
      ],
      "pros": [
        "100% open-source under MIT with 40k+ GitHub stars",
        "Runs entirely in Docker with isolated execution sandboxes",
        "Pairs with local Ollama models or free API endpoints"
      ],
      "cons": [
        "Docker setup required for local containerized orchestration"
      ],
      "date_added": "2026-08-28",
      "scope_caveat": null
    },
    {
      "id": "lovable-dev",
      "name": "Lovable.dev",
      "tagline": "Full-stack software engineer agent turning ideas into real apps with Supabase & GitHub sync",
      "description": "Lovable is a generative web app builder that crafts frontend and backend logic with natural language. Generates clean React code with Tailwind CSS, connects directly to Supabase for authentication and database storage, and pushes commits to GitHub.",
      "website": "https://lovable.dev",
      "category": "appbuilder",
      "subcategories": [
        "Full-Stack",
        "Supabase Sync",
        "GitHub Sync",
        "Web Apps"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free needs no card but limits builds to 5 per day and 30 per month with 20 cloud credits, which fits prototyping rather than shipping.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://lovable.dev/pricing",
        "https://docs.lovable.dev/introduction/subscription-plans"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Automated database and authentication wiring with Supabase",
        "Direct GitHub synchronization with clean git commits",
        "Generous starter credits to build and deploy prototypes"
      ],
      "cons": [
        "Monthly free credit limit applies to heavy iteration loops"
      ],
      "date_added": "2026-09-18",
      "scope_caveat": null
    },
    {
      "id": "aider-chat",
      "name": "Aider",
      "tagline": "Leading open-source AI pair programmer for terminal and git-integrated app development",
      "description": "Aider allows you to build and edit codebases using conversational AI directly from your command line. It edits files across multi-module projects, automatically creates git commits with descriptive messages, and pairs with local Ollama or cloud models.",
      "website": "https://aider.chat",
      "category": "appbuilder",
      "subcategories": [
        "Terminal Agent",
        "Git Integrated",
        "Multi-File Editing",
        "Open Source"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 terminal pair programmer with no account or card; you supply your own API key or local model, so usage is billed by your provider.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/aider-ai/aider",
        "https://aider.chat/docs/config/api-keys.html"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "CLI",
        "MacOS",
        "Linux",
        "Windows",
        "Python"
      ],
      "pros": [
        "Industry benchmark leader in automated git commits and multi-file refactoring",
        "Works offline with Ollama (DeepSeek Coder / Qwen 2.5)",
        "Permissive Apache 2.0 license with active maintainers"
      ],
      "cons": [
        "Command-line interface only (no graphical drag-and-drop)"
      ],
      "date_added": "2026-07-22",
      "scope_caveat": null
    },
    {
      "id": "gpt-engineer",
      "name": "GPT Engineer",
      "tagline": "Specify what you want to build and the AI constructs the entire codebase",
      "description": "GPT Engineer adapts to your style, asks clarifying questions when requirements are ambiguous, and constructs complete multi-file project repositories with documentation, tests, and configuration. The project is now archived, so treat it as a reference implementation rather than an actively maintained tool.",
      "website": "https://github.com/AntonOsika/gpt-engineer",
      "category": "appbuilder",
      "subcategories": [
        "Full Codebase Generation",
        "Open Source",
        "CLI",
        "Project Scaffolder"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed CLI you install yourself; there is no hosted tier and no card, but it needs your own model API key, so you pay for tokens.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/AntonOsika/gpt-engineer"
      ],
      "license": "MIT",
      "platforms": [
        "CLI",
        "Python",
        "MacOS",
        "Linux",
        "Windows"
      ],
      "pros": [
        "Generates complete folder hierarchies and multi-file architectures in 1 run",
        "Asks interactive clarifying questions before writing code",
        "Over 50k GitHub stars with vibrant community extensions"
      ],
      "cons": [
        "Requires careful prompt specification for large production systems",
        "Repository is archived and no longer receives updates"
      ],
      "date_added": "2026-06-15",
      "scope_caveat": null
    },
    {
      "id": "dyad",
      "name": "Dyad",
      "tagline": "Free, local, open-source AI app builder that runs on your own machine",
      "description": "Dyad is an open-source desktop app builder that generates full-stack React applications locally, keeping code and prompts on your computer. It supports bring-your-own model keys or local models via Ollama, offers live preview, and exports clean code with no subscription.",
      "website": "https://www.dyad.sh",
      "category": "appbuilder",
      "subcategories": [
        "Local Builder",
        "Desktop App",
        "Open Source",
        "BYOK"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Free app runs locally with no account or card and gives 20 agent messages a day; code is Apache-2.0 apart from the pro directory; Pro is $20/month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.dyad.sh/pricing",
        "https://github.com/dyad-sh/dyad"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Runs locally with your own keys or local Ollama models",
        "No subscription credits or per-prompt metering",
        "Apache 2.0 with full code export"
      ],
      "cons": [
        "You must supply a model key or run a local model",
        "Smaller template ecosystem than cloud builders"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "onlook",
      "name": "Onlook",
      "tagline": "Open-source visual editor that writes React and Tailwind code into your project",
      "description": "Onlook is a visual-first editor for React apps: click elements in a live preview, prompt a change, and the agent edits your actual codebase with Tailwind classes. The desktop app runs locally against your own repository, so the design and the code stay in sync.",
      "website": "https://www.onlook.com",
      "category": "appbuilder",
      "subcategories": [
        "Visual Editor",
        "React",
        "Tailwind",
        "Open Source"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 editor that the vendor says can be self-hosted free; the hosted cloud is quoted per team with no published free allowance.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.onlook.com/pricing",
        "https://github.com/onlook-dev/onlook"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux",
        "Web"
      ],
      "pros": [
        "Edits real React and Tailwind code instead of exporting mockups",
        "Local desktop app works against your own repository",
        "Apache 2.0 with an active community"
      ],
      "cons": [
        "Cloud version has paid tiers; self-hosting is the free path",
        "Best suited to React projects only"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "ollama",
      "name": "Ollama",
      "tagline": "Get up and running with Llama 3, DeepSeek, and Mistral locally in one command",
      "description": "Ollama allows developers to run open-source large language models entirely locally on macOS, Linux, and Windows. It features a streamlined CLI, built-in REST API, and support for quantized GGUF weights with GPU acceleration across Apple Silicon, CUDA, and ROCm.",
      "website": "https://ollama.com",
      "category": "coding",
      "subcategories": [
        "Local LLM",
        "CLI",
        "API",
        "Self-Hosted"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT app: running models on your own hardware is unlimited and needs no card; the optional cloud is pay-as-you-go per token with 3 concurrent requests.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://ollama.com/pricing",
        "https://github.com/ollama/ollama"
      ],
      "license": "MIT",
      "platforms": [
        "MacOS",
        "Linux",
        "Windows",
        "Docker",
        "CLI",
        "API"
      ],
      "pros": [
        "Zero telemetry and 100% private local inference",
        "One-command installation with instant model downloads",
        "OpenAI-compatible REST API endpoint out of the box"
      ],
      "cons": [
        "Requires substantial unified memory for 70B parameter models"
      ],
      "date_added": "2026-08-15",
      "scope_caveat": null
    },
    {
      "id": "v0-dev",
      "name": "v0 by Vercel",
      "tagline": "Generative UI system to build React, Next.js, and Tailwind CSS components",
      "description": "v0 creates production-ready frontend components, web pages, and full-stack layouts using natural language prompts. It outputs standard React code with Tailwind CSS that can be copied directly into any codebase.",
      "website": "https://v0.app",
      "category": "coding",
      "subcategories": [
        "React",
        "Tailwind",
        "Frontend UI"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free gives new users $5 of one-time credits and 7 messages a day, with no card required; Plus is $30/user/month including $30 of credits.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://v0.app/pricing"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Generates clean TypeScript code using modern Tailwind CSS",
        "Live interactive preview and responsive viewport testing",
        "Generous recurring monthly free generations"
      ],
      "cons": [
        "Rate-limited on complex multi-page iterations on free plan"
      ],
      "date_added": "2026-07-20",
      "scope_caveat": null
    },
    {
      "id": "deepseek-chat",
      "name": "DeepSeek Chat & Coder",
      "tagline": "State-of-the-art open-weights reasoning and code intelligence model",
      "description": "DeepSeek provides high-capability reasoning and coding models with open weights and a completely free web chat and mobile app. There is no paid consumer subscription. Recognized for strong code generation, mathematical reasoning, and low inference latency; the developer API is billed per token rather than free.",
      "website": "https://chat.deepseek.com",
      "category": "coding",
      "subcategories": [
        "Reasoning",
        "Code Gen",
        "Open Weights"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "The consumer chat is free with no paid plan and no card, and the vendor publishes no usage cap; API access is billed separately per million tokens.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.deepseek.com/en/download/",
        "https://api-docs.deepseek.com/quick_start/pricing"
      ],
      "license": "MIT",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "Rivals frontier proprietary models in code benchmarks",
        "Completely free web chat without credit limits",
        "Permissive open-weights release for self-hosting"
      ],
      "cons": [
        "Occasional peak-hour server queue during major releases"
      ],
      "date_added": "2026-09-01",
      "scope_caveat": null
    },
    {
      "id": "continue-dev",
      "name": "Continue.dev",
      "tagline": "Open-source AI code assistant extension for VS Code and JetBrains",
      "description": "Continue is the leading open-source copilot that plugs directly into VS Code and JetBrains IDEs. Connect any local model (via Ollama or LM Studio) or cloud provider for inline autocomplete, codebase chat, and refactoring.",
      "website": "https://continue.dev",
      "category": "coding",
      "subcategories": [
        "VS Code",
        "IDE Extension",
        "Open Source"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 extension you point at your own model or keys; the hosted pricing lists only paid per-token plans, with no free tier to rely on.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://continue.dev/pricing",
        "https://github.com/continuedev/continue"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Works completely offline when connected to Ollama",
        "Full codebase indexing and retrieval augmented context",
        "Zero vendor lock-in; switch models in 1 click"
      ],
      "cons": [
        "Requires configuring model endpoints for best performance"
      ],
      "date_added": "2026-06-11",
      "scope_caveat": null
    },
    {
      "id": "lm-studio",
      "name": "LM Studio",
      "tagline": "Discover, download, and run local LLMs with a polished desktop GUI",
      "description": "LM Studio makes running local open-source models as simple as clicking a button. Browse Hugging Face GGUF models directly within the application, run hardware-accelerated chat, and expose a local OpenAI-compatible server.",
      "website": "https://lmstudio.ai",
      "category": "coding",
      "subcategories": [
        "Desktop GUI",
        "Local LLM",
        "Hugging Face"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "Free desktop app runs local models through llama.cpp and MLX with no card and no usage cap; paid tiers only add cloud models and start at $20/month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://lmstudio.ai/pricing"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Intuitive desktop GUI with automatic GPU offloading detection",
        "Direct search and download from Hugging Face model repository",
        "Built-in local developer server with API testing playground"
      ],
      "cons": [
        "Core application is closed-source freeware, though model weights are OSS"
      ],
      "date_added": "2026-08-01",
      "scope_caveat": null
    },
    {
      "id": "codeium",
      "name": "Devin Desktop (formerly Windsurf & Codeium)",
      "tagline": "Agentic code editor with unlimited Tab completions and inline edits on the free plan",
      "description": "Codeium became Windsurf in 2025 and was renamed Devin Desktop by Cognition in June 2026; codeium.com and windsurf.com now redirect here. The free plan keeps the unlimited Tab completions and inline edits that made Codeium popular, adds a light quota for agent work with a restricted model list, and runs on macOS, Windows, and Linux.",
      "website": "https://devin.ai/desktop",
      "category": "coding",
      "subcategories": [
        "Autocomplete",
        "Agentic IDE",
        "Multi-Language"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free $0 plan has no published numbers, only a light agent quota and limited model access, alongside unlimited tab completions; Pro is $20/month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://devin.ai/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Unlimited Tab completions and inline edits even on the free plan",
        "Full agentic editor with a light free quota for agent tasks",
        "No time limit on the free tier; Pro is optional at $20/month"
      ],
      "cons": [
        "Free plan restricts model choice and caps agent usage"
      ],
      "date_added": "2026-05-18",
      "scope_caveat": null
    },
    {
      "id": "antigravity",
      "name": "Google Antigravity",
      "tagline": "Agent-first IDE, desktop app and CLI from Google, free with a Google account",
      "description": "Google Antigravity is the agent-first development platform that replaced Gemini CLI for individual accounts on 18 June 2026. It combines a desktop app, IDE extensions and the Antigravity CLI on one agent harness that can run several agents in parallel, with scheduled tasks and Gemini 3.8 Flash as the core agent model. The individual plan costs nothing and needs no credit card; usage is metered against a quota that refreshes weekly, and Google does not publish it as a fixed request count.",
      "website": "https://antigravity.google",
      "category": "coding",
      "subcategories": [
        "Agent IDE",
        "CLI",
        "Multi-Agent",
        "Free Tier"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Individual plan is $0 with no card and unlimited tab completions, but the agent quota is described only as meaningful and refreshed weekly, so no figure can be planned around.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://antigravity.google/pricing",
        "https://antigravity.google/docs/plans",
        "https://developers.google.com/gemini-code-assist/docs/deprecations/code-assist-individuals"
      ],
      "license": "Proprietary",
      "platforms": [
        "Web",
        "MacOS",
        "Windows",
        "Linux",
        "CLI"
      ],
      "pros": [
        "No subscription and no credit card for the individual plan",
        "One agent harness across desktop app, IDE extensions and the CLI",
        "Unlimited tab completions on every plan, including free"
      ],
      "cons": [
        "Closed source, unlike the Apache 2.0 Gemini CLI it replaced",
        "Free quota is not published as a fixed number and refreshes weekly",
        "No bring-your-own-key option for extra quota"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "goose",
      "name": "Goose",
      "tagline": "Open-source local AI agent that automates coding and desktop workflows",
      "description": "Goose is an open-source, extensible AI agent that runs locally and connects to the tools you already use through MCP extensions. It writes code, runs commands, manages files, and automates workflows using local models or your own API keys.",
      "website": "https://block.github.io/goose/",
      "category": "coding",
      "subcategories": [
        "Local Agent",
        "MCP",
        "CLI",
        "Desktop"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 local agent with no card; it has no allowance of its own, so you fund it with provider credits or a model free tier.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://block.github.io/goose/",
        "https://block.github.io/goose/docs/getting-started/installation"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux",
        "CLI"
      ],
      "pros": [
        "Runs entirely on your machine with local or bring-your-own models",
        "MCP extension system connects to real tools",
        "Apache 2.0 and backed by Block"
      ],
      "cons": [
        "Configuration is more involved than hosted agents",
        "Requires a model provider or local runtime"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "comfyui",
      "name": "ComfyUI",
      "tagline": "Modular node-based GUI and backend for Stable Diffusion and Flux",
      "description": "ComfyUI is the definitive node-based graphic interface for generative image synthesis. Build modular pipelines for txt2img, inpainting, ControlNet, IP-Adapter, and upscale workflows with precise execution graphs.",
      "website": "https://github.com/Comfy-Org/ComfyUI",
      "category": "design",
      "subcategories": [
        "Image Synthesis",
        "Node-Based",
        "Self-Hosted"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "GPL-3.0 and free for local graph execution with no card; credits only apply to partner nodes and cloud, where five runs are free.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/Comfy-Org/ComfyUI",
        "https://docs.comfy.org/interface/credits"
      ],
      "license": "GPL v3",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux",
        "Docker"
      ],
      "pros": [
        "Unmatched control over generative diffusion graphs",
        "Extremely memory efficient compared to traditional WebUIs",
        "Thriving open-source custom node ecosystem"
      ],
      "cons": [
        "Steeper learning curve for node-based graph composition"
      ],
      "date_added": "2026-07-14",
      "scope_caveat": null
    },
    {
      "id": "recraft-ai",
      "name": "Recraft.ai",
      "tagline": "Generate clean vector SVGs, 3D icons, and brand graphics with daily credits",
      "description": "Recraft is a specialized design canvas that generates editable vector SVG graphics, isometric 3D illustrations, and digital art with consistent brand color palettes. Free tier includes daily generative credits.",
      "website": "https://recraft.ai",
      "category": "design",
      "subcategories": [
        "Vector SVG",
        "Icon Design",
        "Brand Art"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free plan needs no card and allows up to 30 generations a day on daily-resetting credits, with 3 uploads a day and 2 images per prompt, for non-commercial use.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.recraft.ai/docs/plans-and-billing/free-plan",
        "https://www.recraft.ai/pricing"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Exports true scalable vectors (SVG) with clean node paths",
        "Precise style palette matching across illustration sets",
        "Free daily credits are enough for icon and illustration experiments"
      ],
      "cons": [
        "Free-tier images are public and non-commercial; credits are limited"
      ],
      "date_added": "2026-08-20",
      "scope_caveat": null
    },
    {
      "id": "krita-ai-diffusion",
      "name": "Krita AI Diffusion",
      "tagline": "Generative AI painting plugin integrated natively into the open-source Krita canvas",
      "description": "Bring generative inpainting, outpainting, and real-time live drawing directly into Krita. Connects to local ComfyUI backends to paint with AI guidance without leaving the digital painting canvas.",
      "website": "https://github.com/Acly/krita-ai-diffusion",
      "category": "design",
      "subcategories": [
        "Digital Painting",
        "Plugin",
        "Open Source"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "GPL-3.0 plugin that runs locally on your own GPU through a ComfyUI backend; there is no hosted tier, no allowance and no card.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/Acly/krita-ai-diffusion"
      ],
      "license": "GPL v3",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Runs 100% locally with zero cloud dependencies",
        "Interactive real-time canvas preview as you sketch",
        "Preserves layer blending modes and artistic control"
      ],
      "cons": [
        "Requires pre-installed Krita and a local GPU server"
      ],
      "date_added": "2026-06-25",
      "scope_caveat": null
    },
    {
      "id": "invokeai",
      "name": "InvokeAI",
      "tagline": "Self-hosted studio for Stable Diffusion and Flux with a professional canvas",
      "description": "InvokeAI is an open-source generative image studio with a unified canvas, layer-based inpainting, control layers, and workflow support, running on your own GPU. It installs via Docker or Python and exposes a local API.",
      "website": "https://invoke.ai",
      "category": "design",
      "subcategories": [
        "Image Synthesis",
        "Self-Hosted",
        "Canvas",
        "Open Source"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 app you self-host on your own GPU or Apple Silicon; there is no paid tier or card, and hosted GPUs come from third parties.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://invoke.ai/download/"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows",
        "Python"
      ],
      "pros": [
        "Professional canvas with layers and regional control",
        "Runs fully local on your own GPU",
        "Apache 2.0 with a steady release cadence"
      ],
      "cons": [
        "Requires a capable GPU for comfortable use",
        "Initial model setup takes time"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "upscayl",
      "name": "Upscayl",
      "tagline": "Free, offline AI image upscaler for desktop",
      "description": "Upscayl enlarges and restores images locally using Real-ESRGAN and other open models, with batch processing and no uploads. It is a free desktop application for macOS, Windows, and Linux with no account or credit system.",
      "website": "https://upscayl.org",
      "category": "design",
      "subcategories": [
        "Upscaling",
        "Desktop App",
        "Offline"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "Desktop app is free, open source and local with no limits or card; the optional cloud gives a 10-credit trial then charges $24.99 per 300 credits.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://upscayl.org/pricing"
      ],
      "license": "AGPL v3",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Fully offline; images never leave your machine",
        "Batch upscaling with several open models",
        "No account, credits, or watermark"
      ],
      "cons": [
        "AGPL-3.0 license affects redistribution",
        "GPU recommended for large batches"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "claude-ai",
      "name": "Claude.ai",
      "tagline": "Nuanced writing, reasoning, and document analysis on Claude's free plan",
      "description": "Anthropic's conversational assistant, used on a free daily allowance without a card. Renowned for natural human-like prose, strong context comprehension, the Artifacts interactive preview pane, and code assistance. Paid tiers raise message caps and unlock the newest models.",
      "website": "https://claude.ai",
      "category": "content",
      "subcategories": [
        "Writing",
        "Artifacts",
        "Reasoning"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free plan is $0 with no card and includes Projects up to 5, but usage is limited every 5 hours and every week with no message count published.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.anthropic.com/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Exceptional literary quality, tone balance, and nuance",
        "Artifacts pane renders interactive React apps and SVG graphics",
        "Upload PDFs, research papers, and spreadsheets for free"
      ],
      "cons": [
        "Strict message caps per 4-hour window on free accounts"
      ],
      "date_added": "2026-07-01",
      "scope_caveat": null
    },
    {
      "id": "notebooklm",
      "name": "Gemini Notebook (formerly NotebookLM)",
      "tagline": "AI research notebook grounded in your own sources, with Audio and Video Overviews",
      "description": "Renamed from NotebookLM to Gemini Notebook in July 2026, this free Google tool grounds every answer in the notes, PDFs, Docs, and YouTube transcripts you upload. It produces citation-backed summaries, Audio Overviews, Video Overviews, slide decks, and mind maps on a free tier of 100 notebooks and 50 sources each.",
      "website": "https://notebook.google.com",
      "category": "content",
      "subcategories": [
        "Research",
        "Audio Overview",
        "Grounded Notes"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free with a Google account and no card: 100 notebooks, 50 sources per notebook, 50 chats a day, 3 audio overviews a day and 10 reports a day.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://support.google.com/notebooklm/answer/16213268"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "100% free with generous source upload limits",
        "Zero hallucination due to strict document-grounded citations",
        "Groundbreaking Audio Overview creates natural conversational podcasts"
      ],
      "cons": [
        "Limited to analyzing uploaded sources rather than open web search"
      ],
      "date_added": "2026-08-10",
      "scope_caveat": null
    },
    {
      "id": "gemini-flash",
      "name": "Google Gemini (AI Studio free tier)",
      "tagline": "Free API access to Gemini Flash models with multimodal input and long context",
      "description": "Google AI Studio offers a standing free tier on the Gemini API with no credit card. Gemini 3.8 Flash, 3.5 Flash-Lite and other Flash models are listed as free of charge for input and output on the free tier, with multimodal input and long context. Rate limits are set per project and per model and are shown in AI Studio rather than as one published number. Free-tier prompts are used to improve Google products, which is the main reason not to send confidential material through it.",
      "website": "https://aistudio.google.com",
      "category": "content",
      "subcategories": [
        "Multimodal",
        "Free API",
        "Long Context"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "The API free tier is priced at free of charge and needs only a project, but per-model daily quotas are shown inside AI Studio rather than on the pricing page.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://ai.google.dev/gemini-api/docs/pricing",
        "https://ai.google.dev/gemini-api/docs/rate-limits"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "Standing free tier with no billing account required",
        "Flash and Flash-Lite input and output are free of charge on the free tier",
        "High inference speed suitable for real-time applications"
      ],
      "cons": [
        "Free-tier prompts may be used to improve Google products",
        "Pro and flagship models are paid only"
      ],
      "date_added": "2026-06-05",
      "scope_caveat": null
    },
    {
      "id": "chatgpt-free",
      "name": "ChatGPT Free",
      "tagline": "Everyday conversational intelligence, web browsing, and data analysis at no cost",
      "description": "OpenAI's free tier runs its fast Luna model (GPT-6 Luna rolling out from October 2026) with unlimited everyday text chats. Web search, data analysis, file and image uploads, image creation, custom GPTs, and 500 MB of Library storage are included under separate per-tool caps, with no subscription or card required.",
      "website": "https://chatgpt.com",
      "category": "content",
      "subcategories": [
        "Chat",
        "Web Browsing",
        "General Assistant"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free account needs no card and allows unlimited everyday text chats under abuse safeguards, with 500 MB of Library storage; media and research have separate limits.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://help.openai.com/en/articles/9275245-chatgpt-free-tier-faq"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Unlimited everyday text chats on the free tier",
        "Live web search, data analysis, and image creation included",
        "Extensive ecosystem of pre-configured custom GPTs"
      ],
      "cons": [
        "Uploads, images, voice, and deep research are capped per tool",
        "Ads may appear on the free tier in some countries"
      ],
      "date_added": "2026-04-12",
      "scope_caveat": null
    },
    {
      "id": "librechat",
      "name": "LibreChat",
      "tagline": "Self-hosted, multi-model chat UI with agents, MCP tools, and code interpreter",
      "description": "LibreChat is an open-source ChatGPT-style interface that connects OpenAI, Anthropic, Gemini, and local models in one place. It adds conversation search, agents, MCP tools, file handling, and multi-user access, all self-hosted.",
      "website": "https://www.librechat.ai",
      "category": "content",
      "subcategories": [
        "Self-Hosted",
        "Multi-Model",
        "Chat UI",
        "MCP"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed chat UI you self-host with no subscription and no card; it ships without keys, so you add your own provider key or a local model.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.librechat.ai/about",
        "https://www.librechat.ai/docs/quick_start/first_chat"
      ],
      "license": "MIT",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows",
        "Node.js"
      ],
      "pros": [
        "One interface for cloud and local models",
        "Agents, MCP tools, and file search built in",
        "MIT licensed and actively maintained"
      ],
      "cons": [
        "You supply model API keys or a local runtime",
        "Multi-user setups need a Docker deployment"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "open-notebook",
      "name": "Open Notebook",
      "tagline": "Open-source notebook that grounds answers in your own sources, with audio overviews",
      "description": "Open Notebook turns PDFs, links, and notes into a private research notebook with grounded chat, summaries, and podcast-style audio overviews. It runs locally and supports many model providers, so sources never leave your infrastructure.",
      "website": "https://www.open-notebook.ai",
      "category": "content",
      "subcategories": [
        "Research Notebook",
        "Grounded Chat",
        "Self-Hosted",
        "Audio Overview"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed notebook you run yourself in Docker with no card; it needs your own model key or a local Ollama model, and has no paid tier.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.open-notebook.ai/get-started"
      ],
      "license": "MIT",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Private alternative to NotebookLM with local storage",
        "Audio overviews and citation-backed chat",
        "MIT licensed with pluggable model providers"
      ],
      "cons": [
        "Requires Docker plus API keys or local models",
        "Less polished than the hosted original"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "kokoro-tts",
      "name": "Kokoro TTS",
      "tagline": "Open-source 82M parameter text-to-speech model with natural human cadence",
      "description": "Kokoro is a breakthrough lightweight open-source text-to-speech model (82 million parameters) capable of synthesizing expressive, natural speech in English, Japanese, French, and Spanish with minimal compute overhead.",
      "website": "https://huggingface.co/hexgrad/Kokoro-82M",
      "category": "audio",
      "subcategories": [
        "Text-to-Speech",
        "Open Weights",
        "Lightweight"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 weights for an 82M-parameter TTS model you run locally with no card; hosted APIs bill around $0.06 per hour of audio and publish no free quota.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://huggingface.co/hexgrad/Kokoro-82M"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "MacOS",
        "Linux",
        "Windows",
        "Python",
        "CLI"
      ],
      "pros": [
        "Incredible audio quality for an 82M parameter model",
        "Runs faster than real-time on standard laptop CPUs",
        "Permissive Apache 2.0 license for commercial & personal use"
      ],
      "cons": [
        "CLI / Python setup required for custom voice blending"
      ],
      "date_added": "2026-09-12",
      "scope_caveat": null
    },
    {
      "id": "openai-whisper",
      "name": "OpenAI Whisper",
      "tagline": "General-purpose speech recognition and multilingual transcription model",
      "description": "Whisper is an open-source neural net trained on 680,000 hours of multilingual and multitask supervised data collected from the web. It delivers robust automatic speech recognition (ASR) and language translation.",
      "website": "https://github.com/openai/whisper",
      "category": "audio",
      "subcategories": [
        "Transcription",
        "Speech-to-Text",
        "Multilingual"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "MIT licence covers code and weights; it runs locally with no paid tier, credits or card, and small models transcribe on CPU.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/openai/whisper"
      ],
      "license": "MIT",
      "platforms": [
        "MacOS",
        "Linux",
        "Windows",
        "CLI",
        "Docker"
      ],
      "pros": [
        "Industry benchmark for transcription accuracy and accents",
        "Runs 100% offline with zero cloud audio leakage",
        "Wide community tooling ecosystem (whisper.cpp, MacWhisper)"
      ],
      "cons": [
        "Larger models require 4GB+ VRAM for real-time speeds"
      ],
      "date_added": "2026-03-10",
      "scope_caveat": null
    },
    {
      "id": "elevenlabs-free",
      "name": "ElevenLabs",
      "tagline": "Lifelike speech synthesis, emotional AI voices, and sound effects generator",
      "description": "ElevenLabs produces some of the most realistic human voice generation available. The free tier grants 10,000 credits per month (roughly 10 minutes of speech), shared across text-to-speech, speech-to-text, sound effects, voice design, and music, with no credit card required.",
      "website": "https://elevenlabs.io",
      "category": "audio",
      "subcategories": [
        "Voice Synthesis",
        "Sound FX",
        "Dubbing"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free plan is $0 with no card and gives 10k credits a month, about 10 minutes of speech, plus 3 Studio projects and no commercial licence.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://elevenlabs.io/pricing",
        "https://help.elevenlabs.io/hc/en-us/articles/13315218812177-Do-you-offer-discounted-or-free-plans"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "Unsurpassed emotional range and pacing in voice output",
        "Clean audio generation for podcasts, games, and video voiceovers",
        "Monthly free character allowance resets automatically"
      ],
      "cons": [
        "Instant voice cloning requires paid tier",
        "Free-plan audio is non-commercial and requires crediting ElevenLabs"
      ],
      "date_added": "2026-05-02",
      "scope_caveat": null
    },
    {
      "id": "suno-ai",
      "name": "Suno AI",
      "tagline": "Generate complete studio-quality songs with vocals and instrumentation",
      "description": "Suno produces full 2-minute songs with vocals, lyrics, and instrumentals in any genre from jazz to synthwave. Provides 50 free credits every day (equivalent to 10 song generations) on the free tier.",
      "website": "https://suno.com",
      "category": "audio",
      "subcategories": [
        "Music Synthesis",
        "Vocals",
        "Songwriting"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free plan is $0 forever with no card and gives 50 credits a day, enough for about 10 songs, renewing daily; downloads and commercial rights need a paid plan.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://suno.com/pricing",
        "https://help.suno.com/en/articles/2410049"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Daily 50 credits allow regular experimentation without payment",
        "High musicality with diverse instrumentation and vocal styles",
        "Generate from custom lyrics or prompt themes"
      ],
      "cons": [
        "Free-plan songs are personal use only with no commercial rights",
        "Free accounts have no monthly downloads and use the v6-mini model"
      ],
      "date_added": "2026-06-18",
      "scope_caveat": null
    },
    {
      "id": "piper-tts",
      "name": "Piper TTS",
      "tagline": "Fast, lightweight neural text-to-speech that runs on a Raspberry Pi",
      "description": "Piper is a fast local neural text-to-speech engine built on VITS and designed for low-resource hardware. It ships dozens of voices across many languages and runs offline from the command line or Python with near real-time synthesis on a CPU.",
      "website": "https://github.com/OHF-Voice/piper1-gpl",
      "category": "audio",
      "subcategories": [
        "Text-to-Speech",
        "Offline",
        "CLI",
        "Low Resource"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "GPL-3.0 engine installed with pip that runs offline on CPU with no card or paid tier; voice model rights vary by model card.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/OHF-Voice/piper1-gpl"
      ],
      "license": "GPL v3",
      "platforms": [
        "Linux",
        "MacOS",
        "Windows",
        "CLI",
        "Python"
      ],
      "pros": [
        "Faster than real time on a plain CPU",
        "Dozens of voices across many languages",
        "Completely offline with no account"
      ],
      "cons": [
        "GPL-3.0 licensing for redistribution",
        "Voice quality below large cloud models"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "f5-tts",
      "name": "F5-TTS",
      "tagline": "Open-source voice cloning and text-to-speech with natural prosody",
      "description": "F5-TTS is a flow-matching text-to-speech model that clones a voice from a short reference clip and synthesises natural speech in English and Chinese. It runs locally on GPU or CPU and ships with a Gradio interface and CLI.",
      "website": "https://github.com/SWivid/F5-TTS",
      "category": "audio",
      "subcategories": [
        "Voice Cloning",
        "Text-to-Speech",
        "Open Weights"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Code is MIT and the pretrained models are CC-BY-NC; you run it on your own GPU with no card, no hosted tier and no credits.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/SWivid/F5-TTS"
      ],
      "license": "MIT",
      "platforms": [
        "Python",
        "Linux",
        "MacOS",
        "Windows",
        "CLI"
      ],
      "pros": [
        "Convincing zero-shot voice cloning from seconds of audio",
        "MIT licensed with checkpoints on Hugging Face",
        "Local inference keeps audio private"
      ],
      "cons": [
        "GPU strongly recommended for fast synthesis",
        "Only English and Chinese are well supported"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "ace-step",
      "name": "ACE-Step",
      "tagline": "Open-source music generation model for full songs with vocals",
      "description": "ACE-Step is an open music foundation model that generates complete tracks, including vocals and instrumentation, from lyrics and style prompts. It runs locally, supports LoRA fine-tunes, and is fast enough to iterate on a single GPU.",
      "website": "https://github.com/ace-step/ACE-Step",
      "category": "audio",
      "subcategories": [
        "Music Generation",
        "Vocals",
        "Open Weights",
        "Local"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 music model with weights on Hugging Face, self-hosted through pip, Docker or ComfyUI on your own GPU; no hosted tier or card.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/ace-step/ACE-Step"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "Python",
        "Linux",
        "Windows",
        "Docker",
        "CLI"
      ],
      "pros": [
        "Full-song generation with vocals, free and local",
        "Apache 2.0 with LoRA training support",
        "Fast inference compared with similar models"
      ],
      "cons": [
        "Needs a modern GPU and several GB of VRAM",
        "Quality varies by genre and prompt"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "tripo-3d",
      "name": "Tripo3D",
      "tagline": "Generate textured 3D meshes from text prompts or 2D images in seconds",
      "description": "Tripo3D converts text prompts or single 2D images into textured 3D models with downloadable GLB and FBX formats. Free tier includes monthly credits to generate, inspect, and export meshes.",
      "website": "https://www.tripo3d.ai",
      "category": "video3d",
      "subcategories": [
        "3D Modeling",
        "GLB Export",
        "Game Assets"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free plan needs no card and gives 200 credits a month, about 13 models, with 15 downloads a month and non-commercial public output.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.tripo3d.ai/pricing",
        "https://www.tripo3d.ai/product-doc/getting-started/pricing-credits"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Sub-10-second initial 3D mesh draft generation",
        "Clean export to GLB, OBJ, and USDZ for Three.js/Blender",
        "Generous free starter credits"
      ],
      "cons": [
        "Complex topology requires manual retopology for games"
      ],
      "date_added": "2026-07-28",
      "scope_caveat": null
    },
    {
      "id": "wan-2-2",
      "name": "Wan 2.2",
      "tagline": "Open-weights video generation with text-to-video and image-to-video",
      "description": "Wan 2.2 is Alibaba's open-weights video generation family, covering text-to-video, image-to-video, and speech-to-video with a mixture-of-experts architecture. The weights are downloadable and runnable locally through ComfyUI or the reference code.",
      "website": "https://github.com/Wan-Video/Wan2.2",
      "category": "video3d",
      "subcategories": [
        "Text-to-Video",
        "Open Weights",
        "Local",
        "ComfyUI"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 code and weights for 5B and 14B video models, self-hosted on your own GPU; the 14B checkpoint needs large or multi-GPU memory.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/Wan-Video/Wan2.2"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "Python",
        "Linux",
        "Windows",
        "Docker"
      ],
      "pros": [
        "Open weights under Apache 2.0 for local generation",
        "Strong motion quality for an open model",
        "Drops straight into ComfyUI pipelines"
      ],
      "cons": [
        "Large VRAM requirements for full-quality output",
        "Setup is technical compared with hosted tools"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "ltx-video",
      "name": "LTX-Video",
      "tagline": "Fast open-weights video model that generates clips faster than real time",
      "description": "LTX-Video from Lightricks is an open-weights diffusion transformer that produces short video clips quickly, with text-to-video and image-to-video modes. It runs locally and integrates with ComfyUI and Diffusers.",
      "website": "https://github.com/Lightricks/LTX-Video",
      "category": "video3d",
      "subcategories": [
        "Text-to-Video",
        "Fast Inference",
        "Open Weights"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 repository and weights you run locally on an NVIDIA GPU; the separate LTX Studio app gives 800 one-time credits.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/Lightricks/LTX-Video"
      ],
      "license": "Custom / Open Weights",
      "platforms": [
        "Python",
        "Linux",
        "Windows",
        "Docker"
      ],
      "pros": [
        "Generation speed well above most open video models",
        "Runs on consumer GPUs",
        "Active ComfyUI and Diffusers integration"
      ],
      "cons": [
        "Custom open-weights license rather than Apache or MIT",
        "Clip length and resolution are limited"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "hunyuan3d",
      "name": "Hunyuan3D 2.x",
      "tagline": "Open-weights image-to-3D model that produces textured meshes",
      "description": "Hunyuan3D from Tencent converts a single image into a textured 3D mesh with a two-stage shape and texture pipeline. The weights are downloadable, runnable locally from Python, and export to standard mesh formats.",
      "website": "https://github.com/Tencent-Hunyuan/Hunyuan3D-2",
      "category": "video3d",
      "subcategories": [
        "3D Modeling",
        "Image-to-3D",
        "Open Weights"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Tencent community licence, unavailable in the EU, UK and South Korea and needing a separate licence above 1M monthly users; self-hosted with no card.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/Tencent-Hunyuan/Hunyuan3D-2"
      ],
      "license": "Custom / Open Weights",
      "platforms": [
        "Python",
        "Linux",
        "Windows",
        "Docker"
      ],
      "pros": [
        "Free local image-to-3D with downloadable weights",
        "Good texture quality for an open pipeline",
        "Exports usable meshes for DCC tools"
      ],
      "cons": [
        "Tencent community license, not a standard OSI license",
        "Significant GPU memory and setup time"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "meshy",
      "name": "Meshy",
      "tagline": "Text and image to 3D cloud generator with a no-card free tier",
      "description": "Meshy generates textured 3D assets from text prompts or reference images in the browser, with PBR textures and game-engine-ready exports. The free plan includes 100 monthly credits and 10 downloads of untextured Meshy 6 Lite models, with no credit card required.",
      "website": "https://www.meshy.ai",
      "category": "video3d",
      "subcategories": [
        "3D Modeling",
        "Text-to-3D",
        "Cloud",
        "Game Assets"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free plan needs no card and gives 100 credits a month for up to 10 assets, with 10 downloads a month of Meshy 6 Lite only and CC BY 4.0 output.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.meshy.ai/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "100 free credits every month with no card",
        "Fast cloud generation with no GPU needed",
        "Exports GLB, FBX, OBJ, and USDZ"
      ],
      "cons": [
        "Free downloads limited to 10 untextured Meshy 6 Lite models",
        "Free output is CC BY 4.0 and requires attribution"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "perplexity-ai",
      "name": "Perplexity AI",
      "tagline": "AI search engine delivering synthesized answers with real-time web citations",
      "description": "Perplexity combines conversational LLMs with live web indexing to deliver direct answers backed by verifiable source citations. Free tier offers unlimited standard searches and recurring Pro search queries.",
      "website": "https://perplexity.ai",
      "category": "research",
      "subcategories": [
        "Search Engine",
        "Citations",
        "Live Web"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free needs no card and gives 3 Pro searches a day and 1 research query a month, with basic searches effectively unlimited.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.perplexity.ai/hub/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Unlimited quick searches with accurate numbered citations",
        "Focused modes for Academic papers, YouTube, and Reddit",
        "Collections feature organizes research projects neatly"
      ],
      "cons": [
        "Pro searches (deep research) limited on free plan"
      ],
      "date_added": "2026-05-15",
      "scope_caveat": null
    },
    {
      "id": "elicit-research",
      "name": "Elicit",
      "tagline": "Automated literature review and empirical research synthesis for researchers",
      "description": "Elicit searches more than 138 million academic papers and 545,000 clinical trials to answer research questions, extract methodology details, and build literature review tables. The free Basic plan includes unlimited paper search, summaries, and chat with full-text papers, with limited Research Agent and report usage.",
      "website": "https://elicit.com",
      "category": "research",
      "subcategories": [
        "Academic Papers",
        "Literature Review",
        "Synthesis"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Basic is free with no card and includes unlimited search across more than 138 million papers plus unlimited summaries, but only 2 research reports a month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://elicit.com/pricing",
        "https://support.elicit.com/en/articles/8670209"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Extracts sample sizes, methodologies, and outcomes across papers",
        "High research rigor with direct DOI paper references",
        "Free Basic plan with unlimited paper search and summaries"
      ],
      "cons": [
        "Complex multi-column systematic synthesis consumes credits"
      ],
      "date_added": "2026-06-12",
      "scope_caveat": null
    },
    {
      "id": "open-webui",
      "name": "Open WebUI",
      "tagline": "User-friendly, extensible web interface for Ollama and OpenAI-compatible APIs",
      "description": "Open WebUI is the leading self-hosted, offline-first ChatGPT alternative. Connects to Ollama, LM Studio, or external APIs with full chat history, document RAG, image generation integration, and multi-user RBAC.",
      "website": "https://openwebui.com",
      "category": "research",
      "subcategories": [
        "Self-Hosted",
        "Web GUI",
        "Ollama UI"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "BSD-3 community edition is free for unlimited users with no card; it needs your own server and model keys, and a branding clause applies at 50 or more users.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://docs.openwebui.com/license"
      ],
      "license": "Open WebUI License",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Feature parity with ChatGPT Plus in a self-hosted package",
        "Built-in document vector search and retrieval (RAG)",
        "Source-available under the Open WebUI License, actively developed"
      ],
      "cons": [
        "Requires Docker setup to host locally"
      ],
      "date_added": "2026-07-08",
      "scope_caveat": null
    },
    {
      "id": "semantic-scholar",
      "name": "Semantic Scholar",
      "tagline": "Free non-profit academic search across 200M+ papers, with an open API",
      "description": "Semantic Scholar is a research engine from the Allen Institute for AI that indexes hundreds of millions of papers with citation context, TLDR summaries, and author metrics. Web search and the public API are free, with no account needed for basic use.",
      "website": "https://www.semanticscholar.org",
      "category": "research",
      "subcategories": [
        "Academic Search",
        "Citations",
        "Free API",
        "Non-Profit"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "Free account and free API with no paid tier or card; keyed access is limited to 1 request per second over a corpus of 214 million papers.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.semanticscholar.org/product/api"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "Genuinely free, non-profit academic search",
        "Citation graphs and TLDR summaries",
        "Public API for scripts and research tooling"
      ],
      "cons": [
        "No conversational AI layer like newer research assistants",
        "API rate limits apply without a key"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "consensus",
      "name": "Consensus",
      "tagline": "AI research search that answers questions from peer-reviewed studies",
      "description": "Consensus searches scientific literature and synthesises answers backed by peer-reviewed papers, with a consensus meter showing how far studies agree. The free tier includes unlimited paper searches, 10 Pro messages, and 3 Deep reviews per month.",
      "website": "https://consensus.app",
      "category": "research",
      "subcategories": [
        "Evidence Search",
        "Systematic Review",
        "Citations"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free needs no card and gives unlimited paper searches plus 10 Pro messages, 3 deep reviews, 10 study snapshots and 30 API calls a month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://help.consensus.app/en/articles/10087865-subscription-plans"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "Unlimited paper searches on the free tier",
        "Consensus meter summarises agreement across studies",
        "Cites peer-reviewed sources only"
      ],
      "cons": [
        "AI analysis capped at 10 Pro messages and 3 Deep reviews monthly",
        "Full-text access still depends on publisher paywalls"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "n8n-community",
      "name": "n8n Community Edition",
      "tagline": "Self-hosted workflow automation platform with AI agent nodes and LangChain",
      "description": "n8n is a visual workflow automation platform that connects 400+ services with native AI agent nodes, vector stores, and custom LLM prompts. The self-hosted Community Edition is completely free.",
      "website": "https://n8n.io",
      "category": "agents",
      "subcategories": [
        "Workflow Automation",
        "LangChain",
        "Self-Hosted"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "Self-hosted Community edition is free indefinitely with no execution cap under the Sustainable Use License; the cloud free option is only a 14-day trial.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://docs.n8n.io/choose-how-to-use-n8n"
      ],
      "license": "Fair-code (Sustainable Use)",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Unlimited workflow executions when self-hosted on your server",
        "Native AI Agent nodes with tool calling and memory stores",
        "Massive library of 400+ integrations and community templates"
      ],
      "cons": [
        "Fair-code license has restrictions on selling n8n as a hosted service"
      ],
      "date_added": "2026-08-18",
      "scope_caveat": null
    },
    {
      "id": "flowise-ai",
      "name": "Flowise",
      "tagline": "Open-source drag & drop UI to build customized LLM chains and autonomous agents",
      "description": "Flowise provided an intuitive visual canvas to build LangChain workflows, autonomous agents, and RAG pipelines without writing code. The project was archived in August 2026 and reached end of life, so self-hosted installs keep running but no longer receive fixes; the Apache 2.0 code remains forkable.",
      "website": "https://flowiseai.com",
      "category": "agents",
      "subcategories": [
        "Visual Agent Builder",
        "RAG",
        "Drag & Drop"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 core you self-host with your own infrastructure and keys; the repository now states it has been archived, so no new development is expected.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/FlowiseAI/Flowise"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "Docker",
        "Node.js",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Visual node interface makes complex agent logic transparent",
        "One-command npm or Docker installation",
        "Permissive Apache 2.0 license"
      ],
      "cons": [
        "Archived in August 2026 with no further security fixes",
        "Advanced production deployments require custom state management"
      ],
      "date_added": "2026-06-22",
      "scope_caveat": "Scope note: the Flowise repository is archived, so the project is no longer actively developed. The entry is kept because self-hosting still works, but it is a removal candidate."
    },
    {
      "id": "dify-ai",
      "name": "Dify.ai",
      "tagline": "Open-source LLM app development platform combining AI workflow orchestration and RAG",
      "description": "Dify is an open-source development platform that orchestrates LLM applications from prototype to production. Features a visual prompt studio, comprehensive RAG pipeline, and autonomous multi-agent execution.",
      "website": "https://dify.ai",
      "category": "agents",
      "subcategories": [
        "Orchestration",
        "Agent Framework",
        "Enterprise OSS"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Community edition self-hosts free with no licence required; the cloud sandbox gives 200 message credits, 1 member, 5 apps and 5,000 API calls a month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://dify.ai/pricing"
      ],
      "license": "Dify Open Source License",
      "platforms": [
        "Docker",
        "Linux",
        "Web"
      ],
      "pros": [
        "Complete end-to-end platform with user management and logs",
        "Advanced document segmentation and hybrid search RAG",
        "Active open-source community with frequent updates"
      ],
      "cons": [
        "Docker Compose deployment requires 4GB+ RAM on host"
      ],
      "date_added": "2026-07-19",
      "scope_caveat": null
    },
    {
      "id": "crewai",
      "name": "CrewAI",
      "tagline": "Framework for orchestrating role-playing autonomous AI agents in Python",
      "description": "CrewAI enables developers to build collaborative multi-agent teams where each agent has specific roles, backstories, and goals. Perfect for building automated market research, software development, and content pipelines.",
      "website": "https://crewai.com",
      "category": "agents",
      "subcategories": [
        "Multi-Agent",
        "Python Framework",
        "Autonomous Teams"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "The framework is MIT and free to self-host with your own model keys; the hosted AMP free tier allows 2 automations and 50 executions a month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://crewai.com/pricing",
        "https://crewai.com/open-source"
      ],
      "license": "MIT",
      "platforms": [
        "Python",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Clean intuitive Python API for assigning roles and tasks",
        "Seamless support for local Ollama models and cloud APIs",
        "Permissive MIT open-source license"
      ],
      "cons": [
        "Requires Python programming skills to configure"
      ],
      "date_added": "2026-05-24",
      "scope_caveat": null
    },
    {
      "id": "langflow",
      "name": "Langflow",
      "tagline": "Open-source visual builder for AI agents, RAG pipelines, and API-ready flows",
      "description": "Langflow is a low-code visual canvas for building agents, RAG pipelines, and multi-step LLM workflows, with Python code export and an API for every flow. It runs locally via pip or Docker, supports any model provider, and is actively maintained as a successor to Flowise.",
      "website": "https://www.langflow.org",
      "category": "agents",
      "subcategories": [
        "Visual Builder",
        "RAG",
        "Agents",
        "Open Source"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed visual builder with no pricing page; you self-host or run the desktop app with your own compute and model keys, and no card is involved.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/langflow-ai/langflow"
      ],
      "license": "MIT",
      "platforms": [
        "Docker",
        "Python",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "MIT licensed with a very active community",
        "Visual flows export to deployable Python APIs",
        "Works with local and cloud model providers"
      ],
      "cons": [
        "Self-hosting requires Docker or Python setup",
        "Large flows need tuning for production scale"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "activepieces",
      "name": "Activepieces",
      "tagline": "Open-source no-code automation with AI steps and hundreds of integrations",
      "description": "Activepieces is an MIT-licensed automation platform for building event-driven workflows with AI steps, branching, and hundreds of integrations. Self-hosting is free with unlimited flows, and every piece is open source.",
      "website": "https://www.activepieces.com",
      "category": "agents",
      "subcategories": [
        "Workflow Automation",
        "No-Code",
        "Integrations",
        "Self-Hosted"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free forever with no card: 1,000 credits a month, unlimited flows and up to 5 users; the core is MIT and can be self-hosted.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.activepieces.com/pricing",
        "https://www.activepieces.com/docs/about/license"
      ],
      "license": "MIT",
      "platforms": [
        "Docker",
        "Linux",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "MIT licensed with unlimited self-hosted flows",
        "Hundreds of integration pieces",
        "AI steps for LLM-powered automations"
      ],
      "cons": [
        "Cloud plan is paid; self-hosting is the free route",
        "Fewer community templates than n8n"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "appflowy",
      "name": "AppFlowy",
      "tagline": "Open-source, local-first AI workspace and Notion alternative",
      "description": "AppFlowy is an open-source, local-first workspace combining documents, databases, kanban boards, and AI assistants. It runs on your own machine or a self-hosted server, and AI features can use your own model keys.",
      "website": "https://appflowy.com",
      "category": "productivity",
      "subcategories": [
        "Workspace",
        "Notes",
        "Local-First",
        "Self-Hosted"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Personal plan is $0 but capped at 100 MB of storage, 7 MB uploads, one form and one chart, with AI included only as a limited trial.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://appflowy.com/pricing"
      ],
      "license": "AGPL v3",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux",
        "Docker",
        "Web"
      ],
      "pros": [
        "Local-first data with no mandatory cloud account",
        "Docs, databases, and kanban in one workspace",
        "AGPL-3.0 source available for self-hosting"
      ],
      "cons": [
        "AI features need a model key or local model",
        "Some polish still trails Notion"
      ],
      "date_added": "2026-11-14",
      "scope_caveat": null
    },
    {
      "id": "scram",
      "name": "Scram",
      "tagline": "AI web app builder that is free to use and only charges when your app goes live",
      "description": "Scram builds apps by assembling pre-tested component blocks instead of generating every line from scratch, which is how it keeps building free: unlimited apps, no credits and no daily caps, subject to fair use. You pay $9 per published app per month, and changes stay free after launch. Built in London by Scram Software — founders George Bevis (Tide) and Tom Godber (Masabi) — and still in beta, so code export, team access and mobile apps are roadmap items rather than shipped features.",
      "website": "https://www.scram.com",
      "category": "appbuilder",
      "subcategories": [
        "Free To Build",
        "Web Apps",
        "Database",
        "Auth"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Building is $0 with no card, no AI credits and no daily limits for unlimited apps and preview links; publishing costs $9 per app per month.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.scram.com/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Unlimited free building with no credits and no daily caps",
        "Only $9 per app per month once you publish to real users",
        "Changes stay free after launch, so iterating costs nothing extra"
      ],
      "cons": [
        "Publishing, custom domains and collaborators need the paid plan",
        "Beta: no code export, mobile apps or team access yet",
        "Fair-use limits can throttle very heavy AI usage"
      ],
      "date_added": "2026-11-15",
      "scope_caveat": "Scope note: building and previewing apps is free with no card, but publishing costs $9 per app per month. This is a free prototyping environment, not free hosting."
    },
    {
      "id": "design-arena",
      "name": "Design Arena",
      "tagline": "Free blind A/B arena that ranks AI models on real website, app and design output",
      "description": "Design Arena sends one prompt to several frontier models and shows the results head to head with the names hidden until you vote, feeding a public Elo leaderboard. It is free with no paid plans because the preference data is sold to AI labs (built by Arcada Labs, YC S25, $7.9M raised). It is a way to decide which model to pay for, not a builder: there is no code export, and the terms bar commercial use of the service while granting Arcada a broad licence to your prompts and outputs.",
      "website": "https://www.designarena.ai",
      "category": "design",
      "subcategories": [
        "Model Comparison",
        "UI Generation",
        "Benchmark",
        "Free"
      ],
      "pricing_model": "100% Free",
      "pricing_reasoning": "The terms state no fees for using the service and there is no pricing page; it is free with no card and no published usage numbers.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://designarena.ai/terms-and-conditions"
      ],
      "license": "Permissive / Free",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Genuinely free with no plan to upgrade to",
        "Blind comparisons remove brand bias before you commit to a subscription",
        "Covers websites, mobile apps, components, images and more"
      ],
      "cons": [
        "Not a builder: no code export or project handoff",
        "Terms restrict commercial use of the service and outputs",
        "Outputs are public and licensed back to Arcada Labs"
      ],
      "date_added": "2026-11-15",
      "scope_caveat": "Scope note: Design Arena is a model-comparison benchmark, not a tool you build in. It is listed because it helps you choose a model before paying for one — it cannot export code, host a project or hand work off."
    },
    {
      "id": "klaviyo",
      "name": "Klaviyo",
      "tagline": "Email and SMS marketing automation with a permanent free tier",
      "description": "Klaviyo is ecommerce-focused email and SMS marketing with flows, segmentation and sign-up forms. The free plan is genuinely indefinite: 250 active profiles, 500 email sends a month, $5 of mobile messages, $5 of its Composer AI agent, plus social marketing, reviews for 50 orders and 10 helpdesk tickets. It is a public company (NYSE: KVYO), so the tier is unlikely to vanish; sending simply pauses when you pass 250 profiles or 500 sends.",
      "website": "https://www.klaviyo.com",
      "category": "content",
      "subcategories": [
        "Email Marketing",
        "SMS",
        "Automation",
        "Ecommerce"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free plan is permanent, not a trial: 250 active profiles, 500 emails a month, $5 of mobile messages and $5 of Composer AI, with no card.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.klaviyo.com/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "API"
      ],
      "pros": [
        "Free forever: 250 active profiles and 500 email sends a month, no card",
        "Flows, segmentation, forms and reporting are included on the free plan",
        "Public company with a mature deliverability and integration ecosystem"
      ],
      "cons": [
        "Klaviyo branding stays in the email footer and pop-ups on free",
        "No A/B testing until you pay; sending stops at the 250/500 limits",
        "Email support only for the first 60 days; paid email starts at $20/month"
      ],
      "date_added": "2026-11-15",
      "scope_caveat": "Scope note: Klaviyo is an email and SMS marketing platform, and its free Composer AI is one small part of the free plan. It is listed for AI-assisted flows and copy, not as a general-purpose AI tool or model."
    },
    {
      "id": "workshop-ai",
      "name": "Workshop AI",
      "tagline": "Agentic app builder with a free tier, four-model Arena mode and a local desktop app",
      "description": "Workshop builds websites, dashboards, internal tools and AI apps from a prompt, then publishes them to a *.workshop.build URL. Arena Mode has four models build the same prompt so you can compare the results and pick one, and Agent Mode continues from there; the desktop app adds filesystem access and local open-source models. The free Discover plan needs no card and includes publishing, but keeps free app data for only 10 days and grants 250 credits, so it is a prototyping tier rather than a permanent host.",
      "website": "https://workshop.ai",
      "category": "appbuilder",
      "subcategories": [
        "Arena Mode",
        "Cloud + Desktop",
        "Full-Stack",
        "Free Tier"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Discover plan is $0 with no card but the credit allowance is not stated clearly and free app data is kept for only 10 days, so it suits a trial build.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://workshop.ai/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Free Discover plan with no credit card required",
        "Arena Mode compares four models on the same prompt before you commit",
        "Desktop app can run local open-source models for privacy"
      ],
      "cons": [
        "Free plan keeps app data for only 10 days",
        "250 free credits go quickly; paid plans start at $20/month",
        "No refunds on paid plans and unused credits are forfeited"
      ],
      "date_added": "2026-11-16",
      "scope_caveat": null
    },
    {
      "id": "opencode",
      "name": "OpenCode",
      "tagline": "Open-source coding agent for the terminal, desktop and IDE with 200+ model providers",
      "description": "OpenCode is an open-source coding agent that runs in a terminal TUI, a desktop app and editor integrations, and connects to more than 200 model providers as well as local models through Ollama, LM Studio or vLLM. The vendor sells an optional hosted gateway, but the agent works with your own keys or local weights, which is how it stays free.",
      "website": "https://opencode.ai",
      "category": "coding",
      "subcategories": [
        "Terminal",
        "Coding Agent",
        "Multi-Model",
        "Local Models"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed agent you run with your own keys or local models; the hosted free models exist but no daily cap is published, so the reliable free path is your own compute or API key.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://opencode.ai/",
        "https://opencode.ai/docs/models",
        "https://github.com/anomalyco/opencode"
      ],
      "license": "MIT",
      "platforms": [
        "CLI",
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Open source under MIT, so the agent itself is free forever",
        "Reaches 200+ providers plus local models without lock-in",
        "Terminal, desktop app and editor integrations share one agent"
      ],
      "cons": [
        "Free hosted models have no published daily cap",
        "The built-in provider needs an account before it appears",
        "Optional paid gateway costs $10/month and up"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "cline",
      "name": "Cline",
      "tagline": "Open-source coding agent that lives in your editor and terminal and asks before acting",
      "description": "Cline is an AI coding agent for VS Code, JetBrains, Cursor and Windsurf, with a CLI and a desktop app built on the same agent. It supports bring-your-own-key and local runtimes, and every action it proposes needs your explicit approval before it runs, which makes it easier to supervise than fully autonomous agents.",
      "website": "https://cline.bot",
      "category": "coding",
      "subcategories": [
        "VS Code",
        "Coding Agent",
        "Bring Your Own Key",
        "Human Approval"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 extension that is free with your own key or a local runtime; the hosted free models rotate on a limited-time basis with no published quota, so your own credentials are the dependable free route.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://docs.cline.bot/getting-started/free-models",
        "https://cline.bot/faq",
        "https://github.com/cline/cline"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "VS Code",
        "CLI"
      ],
      "pros": [
        "Every tool call and file edit waits for your approval",
        "Bring your own key or run a local model with no key at all",
        "Apache-2.0 codebase with editor, CLI and desktop builds"
      ],
      "cons": [
        "Hosted free models rotate and their quota is not published",
        "Free model usage is unavailable through the Cline API",
        "A $9.99/month pass is needed for higher rate limits"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "kilo-code",
      "name": "Kilo Code",
      "tagline": "Open-source agent for VS Code, JetBrains and the terminal with free model routing",
      "description": "Kilo Code is an agent that runs across VS Code, JetBrains and a CLI, offering more than 500 models, bring-your-own-key and local model routing. Its free tier routes requests to free models through the Kilo Gateway with no card required, and the client code is MIT licensed.",
      "website": "https://kilo.ai",
      "category": "coding",
      "subcategories": [
        "VS Code",
        "JetBrains",
        "Coding Agent",
        "Multi-Model"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed clients that work with your own keys or local models; the free Kilo Auto routing requires no card but the vendor publishes no allowance, so it is treated as open source rather than a generous free tier.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://kilo.ai/pricing",
        "https://kilo.ai/docs/getting-started/using-kilo-for-free",
        "https://kilo.ai/open"
      ],
      "license": "MIT",
      "platforms": [
        "VS Code",
        "CLI"
      ],
      "pros": [
        "Free routing to free models with no credit card",
        "500+ models plus bring-your-own-key and local options",
        "MIT-licensed editor and CLI clients"
      ],
      "cons": [
        "The free routing allowance is not published",
        "Gateway and cloud backend code is source-available, not MIT",
        "Teams plan is $15 per user per month"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "qwen-code",
      "name": "Qwen Code",
      "tagline": "Apache-2.0 terminal coding agent optimised for Qwen models, with any provider you supply",
      "description": "Qwen Code is an open-source coding agent for the terminal, with optional editor integration, tuned for the Qwen model family. The vendor free OAuth tier ended in April 2026, so it is now free software rather than a free service: you run it against your own API key, OpenRouter, Fireworks or an Alibaba Cloud coding plan.",
      "website": "https://github.com/QwenLM/qwen-code",
      "category": "coding",
      "subcategories": [
        "Terminal",
        "Qwen",
        "Coding Agent",
        "Self-Hosted"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "The Qwen OAuth free tier was discontinued on 15 April 2026, so the tool is free software rather than a free service: it works with your own key, OpenRouter, Fireworks or a paid Alibaba Cloud plan.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://github.com/QwenLM/qwen-code",
        "https://qwenlm.github.io/qwen-code-docs/en/users/configuration/auth"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "CLI",
        "VS Code"
      ],
      "pros": [
        "Apache-2.0 agent with no vendor lock-in",
        "Accepts OpenAI, Anthropic and Gemini-compatible endpoints",
        "Custom provider support covers local servers and proxies"
      ],
      "cons": [
        "The vendor free tier ended on 15 April 2026",
        "The route it recommends, Alibaba Cloud Coding Plan Pro, is $50/month",
        "Tuned for Qwen models, so other providers need extra setup"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "mistral-vibe",
      "name": "Mistral Vibe",
      "tagline": "Mistral chat assistant and Apache-2.0 coding CLI in one product, formerly Le Chat",
      "description": "Vibe is Mistral single product for chat and code: the web and mobile assistant formerly called Le Chat, plus a coding agent in VS Code, JetBrains and the terminal. The CLI is Apache-2.0 and can run against your own Mistral API key, while the hosted assistant has a free plan with limited messages, searches and coding sessions.",
      "website": "https://mistral.ai/products/vibe",
      "category": "coding",
      "subcategories": [
        "Chat",
        "Coding CLI",
        "Multimodal",
        "European"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free plan costs $0 with no card and covers chat, search, image editing and limited coding sessions, but Mistral publishes no message or session numbers, so the allowance is limited rather than generous. Pro is $14.99/month for all-day CLI use.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://mistral.ai/pricing",
        "https://mistral.ai/products/vibe",
        "https://github.com/mistralai/mistral-vibe"
      ],
      "license": "Mixed (Apache 2.0 CLI)",
      "platforms": [
        "Web",
        "CLI",
        "VS Code"
      ],
      "pros": [
        "One product covers chat, web search and terminal coding",
        "The CLI is Apache-2.0 and accepts your own API key",
        "Free plan needs no card and includes 100+ connectors"
      ],
      "cons": [
        "Free limits are described only as limited, with no numbers",
        "All-day coding in the CLI needs the $14.99/month Pro plan",
        "Rebranded from Le Chat, so older guides point at the old name"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "kimi",
      "name": "Kimi",
      "tagline": "Moonshot AI assistant with agents, slides, sheets and deep research on a free plan",
      "description": "Kimi is Moonshot AI assistant, available on the web and as a desktop app, bundling chat, agent tasks, slides, sheets, documents, deep research and scheduled tasks. The free plan needs no card and covers small projects with one concurrent agent task, while Kimi Code and larger limits sit behind paid tiers.",
      "website": "https://www.kimi.com",
      "category": "content",
      "subcategories": [
        "Chat",
        "Research",
        "Agents",
        "Documents"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free plan is $0 with no card and states 2 projects, 500 MB project storage, 1 concurrent agent task and 2 scheduled tasks; Kimi Code is excluded, so the allowance is real but small.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.kimi.com/membership/pricing",
        "https://www.kimi.com/"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "Free plan needs no card and includes agents, slides and research",
        "Desktop app alongside the web app",
        "Published free limits rather than vague fair use"
      ],
      "cons": [
        "Free plan is capped at 2 projects and 500 MB storage",
        "Kimi Code is not available on the free plan",
        "Only 1 concurrent agent task and 2 scheduled tasks"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "hugging-face-spaces",
      "name": "Hugging Face Spaces",
      "tagline": "Half a million hosted AI demo apps you can use free, with free static hosting",
      "description": "Spaces is Hugging Face hosting platform and public directory of AI demo apps built with Gradio, Docker or static pages, running on Hugging Face hardware. Browsing and using public Spaces is free within a daily ZeroGPU allowance, static Spaces can be hosted free, while Gradio and Docker Spaces need paid compute to create.",
      "website": "https://huggingface.co/spaces",
      "category": "appbuilder",
      "subcategories": [
        "Hosting",
        "Demos",
        "Gradio",
        "ZeroGPU"
      ],
      "pricing_model": "Limited Free Tier",
      "pricing_reasoning": "Free accounts get 5 minutes of ZeroGPU a day (2 unauthenticated minutes) and can host up to 2 ZeroGPU Spaces, while Gradio and Docker Spaces need a paid plan to create; static Spaces stay free.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://huggingface.co/pricing",
        "https://huggingface.co/docs/hub/en/spaces-zerogpu",
        "https://huggingface.co/docs/hub/spaces-overview"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web"
      ],
      "pros": [
        "Over 500,000 public AI apps to try without an account",
        "Static Spaces can be hosted for free",
        "Free accounts can keep up to 2 ZeroGPU Spaces"
      ],
      "cons": [
        "Most compute Spaces need a paid plan to create",
        "Free ZeroGPU use is capped at about 5 minutes a day",
        "Dev mode and SSH access require a PRO plan"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "jan",
      "name": "Jan",
      "tagline": "Apache-2.0 desktop app that runs local models fully offline with an OpenAI-compatible API",
      "description": "Jan is an open-source desktop app for running local language models on Windows, macOS and Linux entirely offline, with no account and no metering. It downloads models from Hugging Face, exposes an OpenAI-compatible server on localhost, and can also connect to cloud providers if you add your own keys.",
      "website": "https://jan.ai",
      "category": "content",
      "subcategories": [
        "Local Models",
        "Offline",
        "Desktop",
        "OpenAI-Compatible"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "Apache-2.0 desktop app with no account, no paid tier and no metering; any cost comes from hardware or from cloud model keys you choose to add.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://jan.ai/",
        "https://github.com/janhq/jan"
      ],
      "license": "Apache 2.0",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "Runs models fully offline with no account required",
        "Apache-2.0 with no paid tier or usage metering",
        "Serves an OpenAI-compatible API on localhost"
      ],
      "cons": [
        "You supply the model files and the hardware to run them",
        "Cloud models need your own provider keys",
        "Desktop only, with no hosted web version"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "gpt4all",
      "name": "GPT4All",
      "tagline": "MIT-licensed local chatbot that runs on your machine and chats with your own documents",
      "description": "GPT4All is Nomic free desktop chatbot for Windows, macOS and Ubuntu that runs open-source models locally with no cloud calls, no account and no GPU requirement. LocalDocs adds chat over your own files, and the MIT licence allows commercial use without a paid tier.",
      "website": "https://gpt4all.io",
      "category": "content",
      "subcategories": [
        "Local Models",
        "Offline",
        "LocalDocs",
        "Desktop"
      ],
      "pricing_model": "Open Source",
      "pricing_reasoning": "MIT-licensed app with direct downloads, no account, no card and no paid tier; local inference is unmetered and the vendor publishes no cap.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://gpt4all.io/",
        "https://github.com/nomic-ai/gpt4all"
      ],
      "license": "MIT",
      "platforms": [
        "MacOS",
        "Windows",
        "Linux"
      ],
      "pros": [
        "MIT licence permits commercial use at no cost",
        "No account, no card and no cloud requests",
        "LocalDocs answers questions about your own files"
      ],
      "cons": [
        "Desktop only on Windows, macOS and Ubuntu",
        "Larger models need capable local hardware",
        "No hosted fallback if your machine is weak"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "otter-ai",
      "name": "Otter.ai",
      "tagline": "AI meeting notetaker with a free plan of 300 transcription minutes a month",
      "description": "Otter records and transcribes Zoom, Microsoft Teams and Google Meet calls, identifies speakers and writes summaries, on web, desktop and mobile. The Basic plan is free with no card and publishes its limits rather than describing them as fair use: 300 transcription minutes a month, a 30-minute cap on each conversation, three audio or video imports for the life of the account, and the 25 most recent conversations kept open with older ones archived.",
      "website": "https://otter.ai",
      "category": "productivity",
      "subcategories": [
        "Meeting Notes",
        "Transcription",
        "Summaries",
        "AI Chat"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Basic is $0 with no card and states its numbers: 300 transcription minutes a month, 30 minutes per conversation and three lifetime file imports, with 25 recent conversations available. Five hours of meetings a month is enough for real recurring work, not a demo, so it is Generous rather than Limited.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://otter.ai/pricing",
        "https://help.otter.ai/hc/en-us/articles/360047538094-Conversation-import-and-app-limits-on-the-Basic-free-plan"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS",
        "Windows"
      ],
      "pros": [
        "300 transcription minutes a month with no card",
        "Joins Zoom, Teams and Meet automatically and identifies speakers",
        "Live transcription, summaries and AI Chat are included on Basic"
      ],
      "cons": [
        "Each conversation is capped at 30 transcribed minutes",
        "Three audio or video imports in total, not per month",
        "Only the 25 most recent conversations stay open; older ones are archived"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    },
    {
      "id": "fathom-ai",
      "name": "Fathom",
      "tagline": "AI meeting notetaker with unlimited free recordings, transcripts and storage",
      "description": "Fathom records calls, transcribes them and writes an instant summary, with clips, playlists and search across your meeting history. The free plan is genuinely uncapped on the parts that cost money elsewhere: unlimited recordings, transcriptions, storage and downloads, in 38 languages, with bot capture or bot-free capture on Mac. What it limits is the AI layer — advanced summary templates cover the first five calls each month, and AI action items and follow-up emails are paid features.",
      "website": "https://www.fathom.ai",
      "category": "productivity",
      "subcategories": [
        "Meeting Notes",
        "Transcription",
        "Summaries",
        "CRM Sync"
      ],
      "pricing_model": "Generous Free Tier",
      "pricing_reasoning": "Free is $0 with no card and publishes no cap on recordings, transcripts, storage or downloads; the limits sit on the AI layer, where advanced summaries cover the first five calls each month and AI action items are paid-only. Unlimited recording at no cost earns the Generous label despite the summary cap.",
      "last_verified": "2026-11-18",
      "sources": [
        "https://www.fathom.ai/pricing"
      ],
      "license": "Proprietary Free Tier",
      "platforms": [
        "Web",
        "MacOS"
      ],
      "pros": [
        "Recordings, transcripts, storage and downloads are unlimited on Free",
        "Every call gets an instant AI summary at no cost",
        "Clips, playlists and search across meetings are included"
      ],
      "cons": [
        "Advanced summary templates are limited on the free plan",
        "AI action items and follow-up emails require Premium",
        "Bot-free capture is a Mac beta; other platforms use the meeting bot"
      ],
      "date_added": "2026-11-18",
      "scope_caveat": null
    }
  ]
}
