TPU inference for vLLM, with unified JAX and PyTorch support.

Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.

Fine-tuning model on 10-K data across the S&P 500 for financial analyst-grade insights

This repository contains the complete workflow for fine-tuning the Gemma 4 E2B-IT (5.1B) model on the insuranceQA-v2 dataset using Tunix (JAX) on Google Cloud TPUs.

mini cli search engine for your docs, knowledge bases, meeting notes, whatever. Tracking current sota approaches while being all local

Context7 Platform -- Up-to-date code documentation for LLMs and AI code editors

Gemma open-weight LLM library, from Google DeepMind

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

A framework-agnostic, git-native standard for defining AI agents

A Lightweight LLM Post-Training Library

Robust Speech Recognition via Large-Scale Weak Supervision

Spec-driven development (SDD) for AI coding assistants.

Code, prompts, and resources for Gemini CLI: Code & Create with an Open-Source Agent

WeChat-MCP: let Openclaw/Claude/ChatGPT and other AI assistants read and reply to WeChat for you

Open-source MCP server for LinkedIn. Give Claude and any MCP-compatible AI assistant access to profiles, companies, jobs, and messages.

MCP server for Todoist integration enabling natural language task management with Claude

Train transformer language models with reinforcement learning.

τ-Bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains

Public repository for Agent Skills

Research Agent service for the Agentic Workflow course

A Python library for extracting structured information from unstructured text using LLMs with precise source grounding and interactive visualization.

Gerenciador Financeiro Pessoal com Agente de IA (ADK Python, MCP, Gemini, Cloud Run)

Gemini Live Avatar is an open-source web application that aims to provide a conversational, real-time interface using voice, text, and animated avatars in the browser.

CORD: A Consolidated Receipt Dataset for Post-OCR Parsing