Open Source
Our
Projects
We build and maintain open source tools that solve real problems — from AI infrastructure to developer productivity. Each project reflects our commitment to engineering quality and practical utility.
A unified infrastructure for converting between different LLM provider APIs. Seamlessly convert between OpenAI, Anthropic, DeepSeek, Gemini, and more with zero runtime dependencies and full TypeScript support.
- 8 official adapters (OpenAI, Anthropic, DeepSeek, Gemini, etc.)
- Bidirectional conversion with IR pattern
- Zero runtime dependencies
- Streaming & tool calling support
- Desktop app with local proxy server
Open source web editor that supports rich text, block-based, and Markdown editing. Efficient and ready to use out of the box, built on top of ProseMirror and TipTap with Vue 3 support.
- Rich text, block, and Markdown editing modes
- Notion-style editing support
- Extensible plugin architecture
- Built-in light and dark themes
- Multilingual support
Remove watermarks, text, and unwanted objects from images using the LaMa deep learning model. Features a dedicated Gemini watermark removal mode with mathematically precise reverse alpha blending — all processing runs entirely in the browser.
- LaMa deep learning model for intelligent inpainting
- Gemini watermark auto-detection and lossless removal
- 100% browser-local processing — no server uploads
- Multi-step undo/redo with history
- PWA support for offline use
Instantly remove image backgrounds using RMBG-1.4 and U2Net deep learning models. Supports batch processing, real-time before/after comparison, and HD transparent PNG export — fully private, running in your browser.
- Dual AI models: U2Net and RMBG-1.4
- Batch processing with ZIP download
- Real-time comparison slider
- Browser-local processing for privacy
- 5-language internationalization
A powerful browser extension for developers and designers. Inspect elements, pick colors, analyze typography, capture screenshots, and export assets — with keyboard shortcuts and per-tab state management.
- Element inspector with CSS property viewer
- Color picker with magnified preview
- Screenshot tool with watermark options
- Typography and box model analysis
- Available for Chrome, Firefox, and Edge
An open-source e-book chronicling 75 years of artificial intelligence — from the Turing Test to modern LLMs. Approximately 110,000 words across 7 core chapters, systematically tracing every major breakthrough from symbolic AI to deep learning to the Transformer era.
- 7 chapters covering AI's complete evolution
- ~110,000 words of in-depth technical narrative
- From Turing Test (1950) to modern AI agents
- Research-backed with extensive references
- Interactive reading experience with Fumadocs
GitHub trending data for developers and RSS subscribers. Automatically fetches and serves trending repository data via GitHub Actions, providing JSON and RSS feeds filterable by language and date range.
- Daily, weekly, and monthly trending data
- JSON and RSS/XML feed formats
- Filterable by programming language
- Automated via GitHub Actions
- CDN-backed for fast global access
Collaborate
Interested in our work?
We welcome contributions, feedback, and collaboration. Reach out to discuss how we can work together.
Contact Us