Models & repos · most watched this week

Classification · 3,234 likes

A calibrated decision model in the style of Jev. Apache 2.0 license.

Image · 2,097 likes

Text · 1,976 likes

Text · 1,639 likes

Image · 1,516 likes

Text · 600 likes

Vision and text · 3,686 likes

Vision and text · 16,167 likes

Classification · 540 likes

Vision and text · 896 likes

Text · 462 likes

Text · 431 likes

Go · +9,833 stars this week

Alibaba's code review tool: fixed rules plus an AI agent, with line-level comments. Works with OpenAI and Anthropic.

TypeScript · +2,762 stars this week

Anthropic's coding agent that works from the terminal.

Python · +1,895 stars this week

Go · +4,522 stars this week

Tencent's open source platform that turns documents into a queryable RAG, an agent and a self-maintaining wiki.

JavaScript · +6,695 stars this week

A system to tune coding agents: skills, memory and security for Claude Code, Codex, Opencode and Cursor.

JavaScript · +3,867 stars this week

Engineering skills for AI coding agents.

Python · +1,358 stars this week

Anthropic's open source plugins for knowledge work in Claude Cowork.

C · +2,739 stars this week

Runs frontier MoE models on hardware you already own: plain C, no dependencies, experts streamed from disk.

Python · +1,877 stars this week

Tencent Cloud's self-hosted AI assistant, multi-user and multi-agent.

TypeScript · +6,435 stars this week

An environment for running a fleet of coding agents in parallel with your own subscription. Desktop, mobile and remote.

Python · +613 stars this week

A command-line tool to configure and monitor Claude Code.

Rust · +707 stars this week

Cloudflare's implementation of the QUIC protocol and HTTP/3, in Rust.

Kotlin · 5,538 stars since it was created

A phone copilot that reads chats in WeChat, QQ, X or Feishu and suggests replies. Sending stays with the user.

703 stars since it was created

A curated list of tools built for Jev, TypeSafe AI's model for typed decisions.

Python · 503 stars since it was created

A WeChat helper for Windows: local OCR reads the messages, Jev reads the intent and suggests three replies. Sending is always manual.

Python · 437 stars since it was created

Turns any LLM into a Jev-style decision model with real probabilities and no retraining.

Go · 418 stars since it was created

Every agent's model in one place, from the menu bar: Codex on DeepSeek, Claude Code on Kimi.

Python · 376 stars since it was created

A skill for building Marp slide decks: structure, diagrams and layout, with a dark theme and a checking tool.

Python · 344 stars since it was created

A floating WeChat window for macOS: reads the screen, a local model detects intent and risk, and suggests replies. Read-only.

Python · 304 stars since it was created

An agent skill that finds where a photo was taken using maps, elevation, satellite imagery and street view, and shows its work. Works with Claude Code, Codex, Cursor and others.

TypeScript · 303 stars since it was created

A workflow that combines an LLM with Jev.

Python · 233 stars since it was created

A skill for making software launch videos in code, reusing the product's real components: storyboard, music, sound effects and rendering. For Claude Code and Codex.

TypeScript · 211 stars since it was created

Makes your coding agent follow all your project rules.

Python · 172 stars since it was created

Generate realistic teacher documents (ID cards, licenses, letters) for 13 countries via MCP.