Python agents

Training code for advanced RAG techniques - Adaptive-RAG, Corrective RAG, RQ-RAG, Self-RAG, Agentic RAG, and ReZero. Reproduces paper methodologies to fine-tune LLMs via SFT and GRPO for adaptive retrieval, corrective evaluation, query refinement, self-reflection, and agentic search behaviors.

Hierarchical document chunking for RAG. Structure in, structure out.

BleuNova AI Agent: Your ethical, hacker-proof personal AI companion. Self-hosted, autonomous, multi-modal (video, IoT, voice), with built-in continual learning, Grok-powered wit (optional), and zero-trust security — all local-first and open-source.

Agent skill that gives it hands in the browser. 25+ tools to navigate, extract data, execute scripts, intercept APIs — all in user's own Chrome with their login sessions. No passwords needed. 给Agent一双手，像用户一样使用浏览器，25+自动化工具，数据完全本地处理。

The data integrity layer that stops AI agents from silently corrupting shared state — across sessions, time, and concurrent runs.

Token-efficient browser agent for local LLMs — Playwright + accessibility tree + MarkGrab, MCP native.

Local-first observability and control plane for AI agents.

Persistent memory for AI coding agents. Never solve the same bug twice. Works with Claude Code, Gemini CLI, Codex CLI, Kimi CLI.

Knowledge agent for LLMs — online research + local accumulation, gets smarter over time. MCP-ready.

Decentralized agent memory for the Lightning economy. Nostr identity, L402 payments, MCP server.

The flight recorder for AI agents. Tamper-proof, ISO 42001 + HIPAA-compliant audit trails for everything AI agents do on your system. File changes + shell commands + PHI detection + agent identity.

Galactic AI v2.1.0 - Local Voice & Smarter Memory Update. Sovereign, local-first agentic system with 200+ native tools, a Ctrl+K Control Deck, local speech-to-text/TTS, semantic + keyword memory, browser & desktop automation, and a Lite/Full/Custom installer. Zero forced cloud overhead.

A lightweight scaffold for safer, more structured agentic coding — helping developers give coding agents the right project context, guardrails, and workflows without locking into one tool.

Developer acceleration layer for enterprise RAG + agent pipelines. 32 production-style flagships on LangChain, LangGraph, CrewAI, Temporal, OpenAI Agents SDK, pydantic-ai. Permission-aware retrieval, chunk-level citations, schema-enforced output. Built on Model Context Protocol (MCP).

The Agent that evolves through use, distills session experience into domain knowledge, then converts knowledge into executable skills.

Local-first agent debugger with replay, failure memory, smart highlights, and drift detection.

ARI — Autonomous Research Infrastructure: end-to-end autonomous scientific research pipeline

a durable execution framework for Python

An open-source framework that grants AI agents continuous digital life by transforming idle time into self-driven evolutionary loops.

AI Travel Assistant is a powerful AI-powered tool that helps users plan their trips effortlessly. It integrates flight search, hotel booking, and sightseeing recommendations to generate a structured travel itinerary using LangChain, Apify, and OpenAI

RAG chatbot applilcation with fastapi and docker

Public intelligence done by artificial intelligence: an LLM-based OSINT tool designed to perform deep web searches by orchestrating multiple web agents

An open-source, AI-powered chat widget that brings RAG, tool use, and conversational AI directly to your readers. Transform passive browsing into interactive exploration.

A multi-agent system for AI-driven market research, use case generation, and resource collection. It utilizes Langchain, Langgraph, Gradio and Tavily API to research industries, generate AI/ML use cases, and gather datasets and tools. This system helps organizations explore AI solutions and implement innovative strategies effectively.

基于 LangGraph + LangChain + MCP 的政策文本深度分析工具，提供极致的用户交互体验。

利用langgraph搭建agent工作流，采用异步并发及mcp，skill，支持langsmith对agent进行评测及后端服务化

AI-powered Design Authority — 5 specialist agents evaluate architecture decisions and deliver structured rulings in under 60 seconds

Turn Claude Code and Codex from chat assistants into autonomous coding systems

MindBloom AI is an AI-powered mental health voice companion offering empathetic, real-time conversations in 11 Indian languages. Features crisis detection, mood tracking, guided breathing exercises, and therapy booking. Your gentle wellness guide. 🌸

Decentralized Collaboration Framework for Independent AI Agents

SAMAR is an AI-driven tool designed to automate research and analysis processes for mergers and acquisitions (M&A), facilitating the generation of comprehensive reports. By leveraging advanced artificial intelligence techniques, SAMAR streamlines tasks such as data collection, financial modeling, and risk assessment.

Knowledge-Hub open Source project for LLms & Agentic AI

A Model Context Protocol (MCP) server that provides Android TV remote control functionality to AI assistants and other MCP clients.

AI assisted SFC config creation, deployment and operation.

Production-ready Python 3.13+ CLI/API system with Adaptive RAG, multi-engine TTS, OpenRouter key rotation, FastAPI backend, and Next.js dashboard

A curated collection of custom AgentSkills and token-saving MCP servers for extending the capabilities of AI coding assistants like Claude Code.

Local near real-time voice AI assistant

AI Model Pricing Aggregator - Compare AI model pricing across AWS Bedrock, Azure OpenAI, OpenAI, Google Gemini, OpenRouter, xAI and more

Transform complex legal documents into clear, actionable insights with AI-powered analysis.

Benchmark 19 retrieval architectures (vector, contextual, QnA, knowledge graph, LightRAG, hybrid, RAPTOR, PageIndex, BM25, metadata-filtered, temporal, rerank, HyDE, multi-query, late-interaction, SPLADE, agentic, plus two quantum-inspired rerankers) on your own docs. Citable evidence bundles, bootstrap CIs, significance tests.

First learning project on Multi-Agentic RAG using Langchain and LangGraph. You could check it out on the link below 😁

Privacy-first AI middleware with local PII detection, approval gate, and automatic de-anonymization. Query your documents with any LLM — GDPR, KVKK, HIPAA ready.

Better, open-source, LLM wrapper (user management, text processing, file processing, image generation, web search, local models)

OpenAI-compatible multimodal embedding server for Qwen3-VL-Embedding-2B — embed text, images, or both via a simple REST API.

L'intelligence artificielle qui s'éveille à votre métier. Plateforme IA souveraine, hébergée en Europe, auditable. Conscience v4 per-user, 11 plugins, 8 providers LLM, BSL.

Full-stack observability for AI — starting with end-to-end RL post-training, expanding to Agent and inference engine monitoring.

Jarvis is an advanced AI-powered personal assistant that tightly integrates with your desktop system to provide a powerful layer of intelligent automation.

Comprehensive docstring quality vetting for Python projects

MAPLE - Production-ready multi agent communication protocol with integrated resource management, type-safe error handling, secure link identification, and distributed state synchronization.

A workflow tool to track new videos and deliver concise briefings. | 用于追踪新视频并产出精炼简报的工作流工具

ASTRA: Automatic Sizing of Transistors with Reasoning Agents

Capy Video Gen Skill: Multi-shot AI video generation with face identity consistency. 300 experiments, 70% improvement. Script-to-video and idea-to-video pipelines for HappyCapy AI Gateway.

A Retrieval-Augmented Generation (RAG) based question-answering proof-of-concept (PoC) that enables users to query documents using natural language. This system leverages local LLMs via Ollama for enhanced privacy and performance and features a chat-based interface. Built entirely with Python, including both backend an frontend.

A financial chat application powered by LangChain, OpenBB, and Claude 3 Opus.

Automated AI video generation factory. Create faceless YouTube content from prompts with LLM scripts, multi-character TTS, and automated publishing. Built with FastAPI, LangGraph, and Next.js

Filesystem-based persistent memory for LLMs and AI agents -- unlimited context windows without lossy summarization

Open-source AI agent orchestration platform. Multi-agent discussions, visual workflows, 30+ tools, human-in-the-loop governance.

KG-RAG + ToT + multi-agent LLMs for evidence-grounded QA with Neo4j and fine-tuning; reproducible medical case study & evaluation.

多UAV-USV协同巡检系统 - 强化学习调度 + HCA-A* + 2-opt优化

Multi-agent orchestrator with cyclic personal-to-org memory · Apache 2.0 · Python 3.11+

The privacy-first customizable Agentic AI system (and Telegram bot) that researches fragments and remembers every timeline!

AI-powered pharmaceutical supply chain management using LangGraph agents, OpenAI GPT-4, and optimization algorithms.

Open Source AI platform to build Agentic services, ChatGPT Apps and MCP Servers

An Agentic Simulator for Emergency Department Operations

A curated library of machine-readable task prompts for Jules and other coding agents, including prompts for the failures agents actually have: broken setup scripts, vague issues, tests that need services a sandbox cannot start, and pull requests that only read as finished.

Beautiful, themed HTML data displays and research briefings for you and your coding agents.

MCP server for RedNote (Xiaohongshu) — analyze content, comments, and trends with structured tools for LLM workflows.

A complete RAG microservice that ingests from multiple sources, stores embeddings in PGVector with hybrid search, and provides both raw chunk retrieval and LLM-powered answer generation. Built with FastAPI, Celery, Redis, and LlamaIndex. Production-ready with rate limiting, health checks, and Docker deployment.

Persistent memory MCP for AI agents — SQLite, knowledge graph, 13-tool Python MCP + 17-tool BrainBar surface with ToolAnnotations

Self-hosted AI OS: RAG · multi-LLM failover · visual Workflow Studio · Skill Market · multi-tenant workspaces · sandboxed tools · multi-channel (Slack/Discord/Telegram/Feishu) · full observability — one docker compose up

🔥 OSINT (Open Source Intelligence) analysis tool that uses the power of an autonomous AI agent

Macro-factor quantitative signal source for AI agents via MCP. 宏观因子量化信号源。

UK property search UK property portals like Right Move and Zoopla. Normalized JSON output. Built for AI agents.

Self-hosted Compound AI System for sovereign environments. Features token-saving complexity routing, deterministic MCP tools for math/logic, configurable expert templates, causal/reinforcement learning, agentic loops, and a federated Neo4j GraphRAG.

Score the trustworthiness of outputs from any LLM in real-time

Agent plugins for CN and EN journal targeting: candidate discovery, source verification, risk screening, and 2+2 journal decision matrices.

AI Customer Support Agent with RAG + Voice A fully local, privacy-focused AI customer support system powered by open-source models. It supports multi-turn conversations, PDF/manual search (RAG), product recommendations, FAQ reasoning, and voice input/output using local STT + TTS. Built with Mistral 7B GGUF, FAISS, Whisper Tiny, Coqui TTS, FastAPI.

Mission control for AI coding agents

The world's first all-seeing, all-hearing, (almost) all-doing AI agent.

Open-source web search API for AI agents. Tavily-compatible. Apache 2.0, self-hostable on Cloudflare Workers. 10× cheaper.

MCP server that gives Cursor and Claude Code deep Agno framework knowledge — architecture plans, anti-pattern detection, and RAG over the official test suite

Budget-aware context compilation and context firewall for tool-heavy AI agents.

CASCADE is a principled solution to deployment-time learning, which is built on top of the principle of case-based reasoning and utilizes the contextual bandit algorithm to accumulate cases and optimise case retrieval policy over deployment steps.

deepagents-showcase is a hands-on project demonstrating how to build hierarchical, multi-step AI systems using LangChain’s Deep Agents. It showcases task decomposition, planning, sub-agent delegation, and long-running workflows in a practical, real-world setup.

【中文小说写作学习工具】 这是一个帮助热爱写作、想学习写作的个人提升小说创作能力的AI辅助工具。本工具旨在通过人工智能技术，为小说写作爱好者提供学习资源、创作灵感和写作指导。

Official implementation of the paper: "GraphDx: A Cost-Aware Knowledge-Enhanced Multi-Agent Framework for Sequential Diagnosis".

Agentic-DART — autonomous detection & response agent. Architecture-first, not prompt-first. Starts as agentic DFIR; designed to expand toward agentic SOC and beyond.

An AI-powered Multi-Agent Research Assistant built with Generative AI, Agentic AI, LangGraph, RAG, Persistent Memory, FAISS, and FastAPI. It supports multi-format uploads, intelligent PDF analysis, and expert-like Q&A via Researcher, Summarizer, Critic, and Editor agents—offering deep, contextual, and interactive research insights.

Super-Gemini is a local-first, agentic AI system designed to run on Termux combines powerful terminal emulation with an extensive Linux package collection. Android and act as a universal developer + productivity agent.

An intelligent, automated newsletter generation system that crawls the web for AI and tech news, processes it through multiple AI agents, and delivers personalized newsletters via email.

66 curated Claude Code skills for Flutter / Dart development, testing, release, and AI-assisted design-to-code pipelines.

🎯 Optimize LLM token usage by 70-90% with smart context ranking, reducing costs while maintaining quality and performance.

Comparison case study between Gliner V2 vs OpenAI Privacy filter model for PII Redaction

🚀 AI-powered content creation & auto-publishing tool for Chinese social media (小红书/抖音/B站/微博). Generate, schedule, and publish — all from one dashboard.

A complete guide - from zero to hero production AI agents Written for Python developers who have never built an AI agent before.

Sentence-level analysis of 42,971 AI citations across Google AI Mode, Gemini, ChatGPT, Perplexity, Copilot & Grok. Decodes #:~:text= URL fragments to reveal exactly what Google quotes.

Control Spotify from Claude, Cursor, or any MCP client. 100+ tools for playback, playlists, discovery, and curation.

Code once, run everywhere across Ollama, LM Studio, MLX, HuggingFace, vLLM, OpenAI, Anthropic, and more with consistent APIs for tools, structured output, streaming, media, prompt caching, and secure OpenAI-compatible gateway deployment.

AI skill for mirroring public webpages into local pixel-perfect replicas with captured assets, visual QA, and rebranding inventories. 像素级复刻公开网站，内置抓取资源、视觉QA能力和品牌替换清单。