escoffier-labs/brigade
Your agents run loops. Brigade keeps the receipts. Local control plane: share MCP, tools, and memory across harnesses; prove with file receipts; improve only from real exit codes. No daemon, no lock-in.
01 / MACHINE INTELLIGENCE
Models you can run. Agents you can inspect. Tools you can build on.
← All topicsYour agents run loops. Brigade keeps the receipts. Local control plane: share MCP, tools, and memory across harnesses; prove with file receipts; improve only from real exit codes. No daemon, no lock-in.
Open the project to explore the idea and the code.
半调纸拼贴 B-roll 生成 skill:三闸门审批,Gemini Omni Flash 首尾帧组装动画 | Editorial halftone paper-collage B-roll agent skill
UX & AI Product designs skills for Claude, Codex etc.
Vietnamese TTS with instant voice cloning • On-device • Real-time CPU inference • 24kHz audio quality • Chuyển văn bản thành giọng nói tiếng Việt • Text to speech tiếng Việt • TTS tiếng Việt - GitHub - pnnbao97/VieNeu-TTS: Vietnamese TTS with instant voice cloning • On-device • Real-time CPU inference • 24kHz audio quality • Chuyển văn bản thành giọng nói tiếng Việt • Text to speech tiếng Việt • TTS tiếng Việt
Open-source, multi-platform harness for AI agents — a native, multi-threaded operator's console (JVM, not Electron) to run Claude Code, Codex, Gemini or OpenCode with a real browser, terminal, editor, secrets & 100+ MCP tools. Built for enterprises, science & research.
Ten cooperating AI agents that turn long podcasts into short-form videos, automatically. Free and open source, runs on free AI providers.
MCP server + CLI that puts a real human in your coding agent's loop. It publishes work mid-task, a reviewer pins what's wrong and returns a verdict, and the agent iterates until approved.
Run FP8 models on Apple Silicon / MPS in ComfyUI. Compatibility layer for ComfyUI on Apple Silicon.
Defense-in-Depth for AI-Assisted Development.
A desktop application for browsing, searching, and organizing AI-generated images locally. Designed for performance with large collections, focusing on powerful metadata filtering and complete privacy.
Flash Vision-Language-Action Inference for Autonomous Driving
Docker configuration for running VLLM on dual DGX Sparks
Camelid: a Rust-native local inference backend with evidence-gated model compatibility.
Soku CLI — call Soku ads/GA4/PostHog data capabilities from any AI agent or shell.
Based on 4 major Chinese language model platforms, including 8 platform-end versions for web and mobile applications, 614 responses to questions were collected. Each question-platform-end combination was repeated for collection 3 times.
Open-source antivirus for AI agents: block risky tools, secret access, prompt injection, malicious packages, MCP servers, plugins, and skills at runtime.
Local-first conversational AI video editor with a professional multi-track timeline, Agent Skills, MCP integration, and Remotion-powered rendering.
Open-source, backend-free AI video editor for talking-head video — storyboarding, designed graphics, kinetic captions, themes and in-browser WebCodecs export. Drivable by any AI agent over MCP.
An open-source cost controller for AI agent loops — stops a loop when it's actually converged and rolls back before it degrades, instead of running to a fixed max_iterations cap. Real-time loop-gain (Aβ) bands + best-so-far rollback. Adapters for LangGraph, CrewAI, AutoGen, LangChain, OpenAI Agents, and Claude Agent SDK; raw API for custom stacks.
MCP server for the Live Tennis API — give Claude, Cursor and other LLM agents real-time tennis scores, odds and model win-probability
vLLM 0.25.1 serving stack for poolside/Laguna-S-2.1-NVFP4 with DFlash speculative decoding — DGX Spark & RTX 6000 PRO
Open the project to explore the idea and the code.
Open the project to explore the idea and the code.