We check every product on this page once a day and record whether it still answers.
An Apache-2.0, offline-first toolkit that renders in-browser previews of Office, PDF, CAD and other files without a server-side conversion step. For developers embedding document viewers in self-hosted apps.
Watch out · Read-only fidelity varies by file format.
Works offlineOpen sourceSelf-hosted
What to know▼
How it works
A browser-native toolkit that renders read-only previews of Office, PDF/OFD, CAD, archive, email, diagram, 3D, media and data files (206 extensions across 24 pipelines) without a mandatory server-side conversion step.
What's different
Apache-2.0 licensed and offline-first; components can be self-hosted and used across Vanilla JS, React, Vue, Svelte and jQuery.
Best for
Developers embedding document/file preview into private or internal web apps.
We check every product on this page once a day and record whether it still answers.
An open-source toolkit for building AI features — chat, RAG, speech, vision — that runs entirely in the browser with no servers and no data leaving the device. Built for developers.
Works offlineData stays with youOpen source
What to know▼
How it works
An open-source toolkit for running AI entirely in the browser: LLM chat across 76 models, RAG and vector search, Whisper and Kokoro speech, and vision, with no server or API key required. Runs on WebGPU with a WebAssembly fallback.
What's different
Data never leaves the device, and it ships a zero-dependency core plus a shadcn UI registry of 107 components and 36 installable blocks for building on top of it.
Pricing
Open source (MIT license).
Best for
Developers building browser-based AI features without standing up servers or managing API keys.
We check every product on this page once a day and record whether it still answers.
A desktop coding harness built specifically around open-weight models like DeepSeek, Qwen and Kimi, instead of bolting them onto a Claude or GPT-first interface as an afterthought.
Watch out · Depends on local tools like Ollama and MCP — offline/no-cloud setup requires configuring those dependencies yourself.
FreeWorks offline
What to know▼
How it works
A desktop AI coding harness built around open-weight models (DeepSeek, Qwen, Kimi, MiniMax) instead of Claude/GPT, using a structural + lexical code-retrieval engine (Blueprint) that builds a real dependency graph of the codebase rather than embeddings or a vector DB.
What's different
Built specifically for open-weight models' weaker context discipline, instead of treating them as an afterthought behind a Claude/GPT-first interface.
Pricing
Free.
Best for
Developers who want to use open-weight models like DeepSeek, Qwen, or Kimi for coding with retrieval built for their limitations.
Watch out
Depends on local tools like Ollama and MCP — offline/no-cloud setup requires configuring those dependencies yourself.
We check every product on this page once a day and record whether it still answers.
A local-first, self-hostable, open-source notes app with semantic search, autonomous research agents, and voice notes, where data never leaves the user's machine unless they choose to share it.
Watch out · Requires self-hosting, so it needs technical setup compared to a hosted notes app.
Works offlineData stays with youOpen source
What to know▼
How it works
Stores notes locally and adds semantic search, ReAct agents, cited deep research, voice notes, and financial analysis on top, all self-hosted and MIT licensed.
Best for
Technical users who want an AI-augmented notes app without sending their data to the cloud.
Watch out
Requires self-hosting, so it needs technical setup compared to a hosted notes app.
We check every product on this page once a day and record whether it still answers.
A local-first macOS knowledge base for browsing notes, PDFs, and webpages with source-cited Q&A.
Watch out · macOS only.
Works offline
What to know▼
How it works
A local-first personal knowledge base for browsing notes, PDFs, webpages, images, and code, answering questions with source context, and saving useful answers back to ordinary files.
What's different
Can delegate heavier follow-up work to Claude Code or Codex when needed.
