Open-source AI tools worth installing

  1. Cherry StudioCherryHQ/cherry-studio★ 52.9kAGPL-3.0TypeScript

    A cross-platform desktop AI client that puts many cloud models and local Ollama in one interface, with a built-in knowledge base, MCP, agents and 300+ assistant templates.

    Built by a Chinese team with complete Traditional and Simplified Chinese UIs. Supports Chinese models like DeepSeek, Qwen and Kimi as well as international ones; easy to set up.

    InstallDownload the installer from cherryai.com or GitHub Releases
  2. Upscaylupscayl/upscayl★ 50.9kAGPL-3.0TypeScript

    A desktop AI image upscaler that enlarges low-res photos or illustrations 2–16× and restores detail, with batch processing.

    Drag and drop, no Python needed; uses Vulkan so it runs on most GPUs. A free alternative to Topaz Gigapixel.

  3. nanobotHKUDS/nanobot★ 49.3kMITPython

    An ultra-lightweight personal agent framework from the University of Hong Kong's Data Intelligence Lab, providing memory, tools, MCP and chat-app integrations in a small amount of Python.

    Far less code than OpenClaw, so it's easy to read and modify. Good for self-hosting or learning agent architecture; built by a Hong Kong team.

  4. 思源筆記 SiYuansiyuan-note/siyuan★ 46.9kAGPL-3.0TypeScript

    Block-based, local-first note-taking app with backlinks, database views and self-hosted sync, plus built-in AI and MCP so agents can read and write notes.

    Built by a Chinese team with strong Chinese typesetting and input-method support, fully self-hostable. A Chinese-friendly alternative to Notion and Logseq.

    or docker run b3log/siyuan
  5. Janjanhq/jan★ 45.1kCustom licenseRust

    A ChatGPT-style desktop app that runs fully offline, with built-in model downloads and the llama.cpp engine; can also switch to cloud APIs.

    The friendliest local LLM option for people who'd rather avoid the terminal. Small Tauri build, and it doubles as a local API server.

  6. Open Notebooklfnovo/open-notebook★ 40.5kMITTypeScript

    An open-source NotebookLM: import PDFs, web pages and videos for Q&A and summaries, and generate multi-speaker podcasts.

    Self-hostable with any model (including local), so your data doesn't go to Google. Podcast speaker count and voices are customizable.

    InstallStart with Docker Compose per the README (image lfnovo/open_notebook)
  7. A local proxy that lets Claude Code and other agents use DeepSeek, Kimi, GLM, Gemini, OpenRouter or local models, routing automatically by task.

    Maintained by a Chinese developer; especially useful for users who can't access Anthropic directly or want lower costs. A desktop version is available.

  8. Khojkhoj-ai/khoj★ 37.8kAGPL-3.0Python

    A self-hosted AI second brain that searches and answers questions over your Markdown, PDF, Notion and Obsidian data, with scheduled automations and deep research.

    Accessible from Obsidian, Emacs, the browser, mobile and more, with cloud or local models — one service covering many note sources.

  9. Continuecontinuedev/continue★ 36.3kApache-2.0TypeScript

    Chat, inline completion and an agent mode in VS Code and JetBrains, with models and rules defined in config files you control.

    One of the few open-source options with full JetBrains support, and it can run entirely offline on local Ollama models.

    JetBrains: search Continue in the JetBrains Marketplace
  10. Handycjpais/Handy★ 34.5kMITRust

    Press a hotkey and speak; local Whisper/Parakeet models transcribe your speech and type it into any app.

    Fully offline and free — an open-source alternative to paid dictation tools like Superwhisper and Wispr Flow. Supports Chinese.

    InstallDownload the macOS/Windows/Linux installer from GitHub Releases
  11. Serenaoraios/serena★ 30.7kCustom licensePython

    An MCP toolkit built on language servers that lets agents read and edit code by finding symbols, references and functions instead of reading whole files.

    Saves tokens and reduces bad edits in large projects. Works with Claude Code, Codex, Cursor and JetBrains.

  12. Kilo CodeKilo-Org/kilocode★ 27.7kMITTypeScript

    A coding agent as a VS Code/JetBrains extension and CLI, with Architect, Code, Debug and other modes; bring your own API key.

    Roo Code was archived in 2026; Kilo Code is the actively maintained continuation of that lineage, and it works for JetBrains users too.

  13. IndexTTSindex-tts/index-tts★ 24.7kCustom licensePython

    Zero-shot speech synthesis open-sourced by bilibili: clone a voice from a few seconds of reference audio, with control over emotion and exact duration. Strong in Chinese and English.

    Lets you set speech length, which suits lip-synced video dubbing; handles Chinese pronunciation and polyphonic characters better than most open-source TTS.

  14. Translates videos from one language to another: speech recognition, subtitle translation and dubbing, merged back into the video. Can also do transcription or TTS alone.

    Offers a no-install packaged build for Windows, works with local Whisper and many translation/TTS services, and is very actively maintained.

    InstallWindows: download the prebuilt package; macOS/Linux: git clone, then run uv run sp.py
  15. ccusageccusage/ccusage★ 19.2kCustom licenseRust

    Reads local Claude Code, Codex and similar logs and reports token usage and estimated cost by day, month and session.

    Runs with one command, no install. Lets subscribers see their usage and 5-hour limit consumption. Now rewritten in Rust.

  16. VideoLingoHuanshere/VideoLingo★ 18.9kApache-2.0Python

    One click for speech recognition, sentence segmentation, translation, subtitle alignment and AI dubbing, outputting Netflix-standard single-line subtitles.

    Uses an LLM to understand before translating, then revises, so subtitles beat sentence-by-sentence machine translation. Good for turning foreign-language videos into Chinese subtitles.

  17. Nanobrowsernanobrowser/nanobrowser★ 14.2kApache-2.0TypeScript

    Multiple AI agents operate web pages in your Chrome — comparing prices, filling forms, gathering data — using your own LLM API key.

    Similar to OpenAI Operator or Comet, but fully open source and running in your local browser, so your logged-in sessions never go to a cloud service.

    InstallChrome Web Store: search “Nanobrowser”
  18. Inserts translations under each paragraph of a web page for side-by-side bilingual reading, and also translates input boxes, subtitles and selected text. Works with DeepL, Google, OpenAI-compatible APIs or local Ollama.

    Minimal UI and customizable rules, available as both an extension and a userscript. An open-source alternative to Immersive Translate's paid features, with Traditional Chinese output.

    InstallSearch “KISS Translator” in the Chrome Web Store, Microsoft Edge Add-ons or Firefox Add-ons, or install the userscript
  19. Call AI from a floating window on any page: translate selections, summarize, show answers next to search results. Supports many APIs and web-app accounts.

    Long-running and still maintained, with support for Chrome, Edge, Firefox and Safari. Good for anyone who wants a general-purpose AI toolbar.

    InstallSearch “ChatGPTBox” in the Chrome Web Store, Microsoft Edge Add-ons or Firefox Add-ons
  20. Uses an LLM for bilingual web page translation and explains words, phrases and grammar at your level, turning foreign-language browsing into language learning.

    New in 2025 but updated frequently. Puts translation and language learning in one sidebar; bring your own OpenAI, DeepSeek or other API key.

    InstallChrome Web Store: search “Read Frog”

Updated Oct 25, 2026 · ★ GitHub