Python agents

LLM 驱动的一体化图片对话插件。自动生成多维度图像标签，并通过多模态语义分析，实现智能斗图、主动表情、表情包分享等多种功能，支持 4 种导入图库方法。

An AI-powered chatbot system using RAG (Retrieval-Augmented Generation) to help PokeMMO players. Built with a custom data processing pipeline that prioritizes control, performance, and Vietnamese language support.

DocBuddy.ai 🩺 Your Personalized 🪄 Doctor Buddy 👨🏻‍⚕️

A modern, fully-async framework for Bale (بله) bots — inspired by aiogram.

在 AstrBot 中统一管理 Palworld 单服/多服，让群友通过状态、日报、事件与排行参与服内日常；提供可视化设置、分组/逐命令权限，以及默认关闭、仅管理员可用的服务器管控。基于官方 REST API。

Lightweight Python client for DuckDuckGo's AI Chat API (unofficial duck.ai). Supports GPT-4o, Claude, LLaMA, Mixtral & more. Fast, minimal, production-ready.

🧠 Happy2Help is your mental health chatbot

Local-first eval harness for unstructured-document extraction — compare LLMs, OCR/IDP tools, and strategies (cascade/ensemble/verify) on the same cohort.

Long-horizon agent execution harness — reliable autonomous runs for Claude Code, Codex, OpenHands, and custom agents. Goal graphs, spin detection, HITL gates, fork/merge, 8 strategies, 6 validators.

Full-stack WhatsApp inbound SDR system with AI — lead qualification, CRM integration and calendar scheduling

TEKION Slide Generator v4 — AI-driven presentation slide generator (OpenAI GPT-image-2 / Gemini 3.1 Flash). Markdown → 16:9 2K slides → PPTX/PDF. Provider abstraction, Visual/Balanced styles, logo integrity, reference image support.

Graxon: Open-Source Agentic Hybrid RAG to eliminate hallucinations through a persistent Knowledge Graph layer.

A production-style retrieval-augmented generation (RAG) pipeline for arXiv research papers — fetch new papers daily, parse them with a structure-aware PDF model, store metadata in Postgres, chunk + embed each paper, and serve hybrid (BM25 + vector) semantic search via OpenSearch.

AI-powered Toutiao content generation, creator activity discovery, automation and protocol publishing dashboard.

🤖 Comprehensive solution for running Ollama/vLLM API servers in Kaggle notebooks with OpenAI-compatible endpoints

AI-driven analysis workflow that uses LLMs to analyze narratives, content patterns, and information structures.

A semantic type system for validating LLM output intent — Pydantic for meaning.

This repository contains an intelligent web scraping solution that uses ScrapeGraphAI for LLM-powered content extraction and LangGraph for orchestrating the scraping workflow. The system can intelligently crawl websites, extract content using natural language instructions, and search for specific information.

Build enterprise AI agents with DAG scheduling, hierarchical memory, 20+ tools, and multi-level permissions for safe local or sandboxed execution

An intelligent robotic task planner built on ROS 2 Humble, leveraging Large Language Models for high-level reasoning and Nav2 for autonomous navigation

Benchmarking tool for stress testing self-hosted LLMs, Azure OpenAI, and OpenAI APIs. Measure latency, throughput, and optimize performance. 🚀

Turn clothing photos into Vinted listing drafts with AI-assisted copy and comparable-based pricing.

AI-powered market intelligence system for traders — combines real-time news, strategy analysis, macro awareness, and automated reporting using Claude Flow, GPT, and financial APIs.

One-file Python SDK for OpenAI-compatible LLM APIs: streaming + reasoning, tools from callables, strict structured output, vision/audio/video/files.

Git LFS for your LLM context — 20x Context Savings — an MCP server that keeps big documents out of the model and returns byte-exact quotes for a fraction of the tokens.

Source-backed skills that turn official Claude and Codex training into reliable daily workflows.

Claude skill: 把一个生活妙招选题端到端做成可直接发布的竖屏 AI 短视频(自带 AI 配音)。双锚定一致性 + 全中文分镜 + 火山 ARK 单网关。End-to-end AI short-video skill with native voiceover.

每日 AI 大厂动态卡片：抓取 Anthropic 等官方动态→中文策展→小红书/公众号卡片→GitHub Pages 审核页。内含可安装 Agent Skill ai-daily-digest。

💼 The portable memory layer for AI coding tools — one shared brain that follows you across Claude Code, Cursor, Windsurf & Conductor. Install once, never re-explain your project again.

Python implementations, C++ structures, and CUDA GPU kernels that demonstrate the core algorithms, memory allocators, quantization schemes, disaggregation transports, and latency profilers used in modern LLM serving engines like vLLM, SGLang, TensorRT-LLM, llama.cpp, NVIDIA Dynamo, and LMCache.

RAG (Retrieval-Augmented Generation) backend for organisational Terraform module knowledge.

Singapore job search and resume workflow with 76K+ live postings, RAG matching, ATS scoring, and validation-gated AI tailoring.

RAG answers you can verify — citations, faithfulness checks, exportable evidence reports

Local RAG semantic search for Obsidian vaults — Ollama + LanceDB + MCP. Search your markdown vault by meaning, not keyword.

本地法律知识库建设一体化工厂 — 从 PDF/DOCX/图片到可对外服务的业务咨询智能体，一个包搞定

Automated Medical Content Generation using n8n, LLM prompts, and SEO workflows.

Fast markdown RAG with multi-provider embeddings (Vertex AI, Gemini, OpenAI, Voyage), incremental indexing with mtime/size fast-path, stale vector pruning, Milvus vector store.

MetaRAG is an open-source engine that takes the guesswork out of RAG pipeline design. Instead of manually tuning chunking strategies, retrieval backends, and rerankers, MetaRAG benchmarks them against your own queries and routes each one to whichever configuration actually performs best on your data.

Local-first Plaud recording pipeline and personal knowledge timeline with Python, SQLite, Qdrant, RAG, 3D graphs, MCP tools, and a SwiftUI companion.

Trợ lý AI tiếng Việt chạy hoàn toàn offline trên máy cá nhân — giọng nói, bộ nhớ dài hạn, RAG an ninh mạng có trích nguồn, giao diện orb kiểu Jarvis. Tự fine-tune QLoRA trên Qwen3-4B, không phụ thuộc API bên ngoài.

Local-first content research pipeline and Creator Intelligence System

CI-friendly RAG evaluation: retrieval metrics, LLM-as-judge faithfulness, and a regression gate that fails your build when quality drops

Minimalist, deterministic, lightweight, fast, and portable verification of AI-generated text: measure groundedness in a source as verbatim reuse (contiguous matching) or verbatim paraphrasing (LCS). Use as a decorator over generate(), a CLI or a CI gate. Useful for research and production purposes.

AI-powered educational assessment generation platform for instructors, trainers, and enterprise knowledge management systems.

agent-skills for chonkie's ecosystem

QAFD-RAG is a graph-based retrieval-augmented generation framework that uses query-aware flow diffusion to retrieve contextually relevant subgraphs from a knowledge graph with statistical retrieval guarantees.

HackerRank Orchestrate hackathon AI agents: a multi-modal damage-claim verifier (3rd of 1,773) and a support-triage RAG agent (60th of 1,349). The model describes; plain code decides

Sync Canvas courses into a local Obsidian vault — transcribed lecture notes, a cross-lecture concept graph, homework and deadlines — then study it with Claude or Gemini over MCP.

The missing MCP gateway you always wanted. Elevate MCP servers into authenticated HTTP endpoints. Self-hosted, in one container.

Turn your AI coding history (Claude Code, Cursor, Copilot, ChatGPT exports, git commits) into a versioned, reviewer-ready résumé — 9 locales, JIS履歴書 grid, Europass, reviewer audit. For the vibe coding era.

Spatial Memory replaces standard cosine-similarity RAG with physics-inspired 8D manifold where memories and beliefs have: Mass — derived from confidence and affective encoding ; Temperature — recency heat that cools over time (Lorentzian decay); and Gravity — `F = T × m / d²` — hot, massive, nearby concepts pull attention

This project is a secondary development based on CoreCoder, with feature upgrades.

Momo — our first attempt at a digital store manager, and the start of this journey. Now carried forward as Saros (in preparation).

Link to insight - AI distillation from any URL

AI analyst toolkit for Terra Invicta: save analyzers, code-verified mechanics docs, and a lessons library

IRCv3 platform adapter for Hermes Agent (SASL, observe-mode, runtime channel agency, chathistory persistence)

CLI agent that fills out online job application forms - cover letters, screening answers, Greenhouse/Lever/Workday - using AI-driven browser automation.

🤖 Jarvis — Ein hochmoderner, deutschsprachiger KI-Assistent basierend auf Gemini Live. Mit nativer Sprachausgabe, MCP-Unterstützung und Vision-Features.

面向电赛视觉的泰山派 RK3566 Linux Codex/CC Skill：从题目分析、板端探测到 OpenCV/V4L2、UART/GPIO/RKNN 验证与工程验收。

不動産鑑定士が公開する白箱方式の土地査定OSS。MLIT取引データからヘドニック回帰係数を都度推定、業者用シート（係数全開示）と顧客用シートを生成。Claude Coworkerシリーズ。White-box AVM skill for Japanese land price assessment. Re-estimates hedonic coefficients on every run from MLIT transaction data, generating broker-side worksheets (full coefficient disclosure) and client-facing assessment reports (xlsx). Part of the Clau

Evidence-first Agent Skill for economics and management paper writing, polishing, peer review, citation integrity, method safety, LaTeX, and auditable revision.

Hermes WeChat Multi-Account Plugin (iLink Bot API)

A skill that analyzes A-share trading signals using chip structure, resonance, KDJ, Wyckoff, and Chan theory. 利用筹码结构、共振、KDJ、威科夫和缠论分析A股信号的skill。

让任何 AI 真正"看"视频：纯本地的视频阅读 skill——faster-whisper GPU 转写 + ffmpeg 场景感知抽帧，双通道带时间戳（t=MM:SS），数据不出机。1 小时视频 3~5 分钟读完。Give any AI agent eyes for video: 100% local video-reading skill — faster-whisper GPU transcription + scene-aware ffmpeg frame extraction, dual-channel with t=MM:SS citations. Nothing leaves your machine.

Voice teammate for Claude Code: talk to your live session hands-free, hear replies spoken. Local whisper STT + macOS TTS. /voice-on in any session.

Persian-language natural-language-to-SQL for enterprise data warehouses. AST-validated SQL against a closed allowlist, per-analyst column ACLs, full audit trail, multi-turn sessions, and a two-role admin panel. Runs against any OpenAI-compatible endpoint - local or hosted.

One-click Render deploy template for nanobot — a lightweight self-hostable AI agent

OpenAI-compatible media generation gateway for AI Agents. Supports multi-provider image/video generation, model routing, prompt enhancement, and built-in Web Studio.

NoCodeWork Growth OS — open-source Growth OS for any AI agent: point it at a business URL, persist USP/UVP/ICP context, run a growth audit (SEO, speed, GEO/AI-visibility, competitors, live GA4/GSC), then delegate to marketing skills. Runs on Claude Code, Codex, Hermes & MCP.

AI Agent Skills library to facilitate technical lead in a software project

Claude Code skill for local LinkedIn lead scouting and outreach automation.

Natural-language-to-SQL agent built with LangGraph and FastAPI. Ask questions in plain English, get answers from your database. Powers QueryDB (querydb.in) — self-host or use free at querydb.in

基于 Dify + Ollama + MCP 实现的全本地化电商客服智能体

Code for 'Evaluating Collective Behaviour of Hundreds of LLM Agents' (AAMAS SE 2026)

Build self-evolving AI agent harnesses with portable harness units, artifact-aware testing, trace-backed diagnosis, and evidence-gated promotion.

LangGraph and LangChain retriever, tools, and navigator subgraph for OKF (Open Knowledge Format) bundles.

Production-grade prompt engineering curriculum — 7 modules covering foundations, reasoning, RAG, security, and evaluation. Model-agnostic (OpenAI/Anthropic/Ollama), runnable notebooks, mini-projects, Mermaid diagrams. Not prompt tips — prompt engineering as a discipline.

Autonomous multi-agent AI system for intelligent GitHub PR review and bug detection. Built with LangGraph, RAG + tree-sitter, and GitHub MCP for deep codebase understanding. Automatically raises issues, drafts fixes, and opens PRs as smart-pr-review-bot[bot]. Supports Review Only, Human-in-the-Loop, and Auto-Pilot modes with full LangSmith tracing.

AI-powered ConstructionRetrieval-Augmented Generation system built with FastAPI, Streamlit, Groq LLMs.

OpenGPT, An open AI workspace for conversations, knowledge, and productivity.

A governance framework for multi-agent AI systems. Structure, lifecycle, and quality controls for AI agent orchestration.

A virtual life form dedicated to progress together with you and your team / 一个致力于与您和您的团队共同进步的虚拟生命体

OpenCode + Ollama multi-agent workflow with builder/reviewer agents, headless runs, skills, validation gates, local session evidence, and redaction tools.

🤖 Enhance coding with Claude and Codex, a library for duo peer programming that combines AI strengths for superior code generation and reviews.

Production multi-agent Claude Code skill — turns a voiceover MP3 into a full whiteboard-explainer YouTube episode. Brief-on-disk + SHA256 verify, gated 2-wave generation, vision review loops.

Your AI on rails, on your terms — an open-source Claude Code plugin that enforces every process as it happens, so your AI can't fake the work. Free, on the subscriptions you already pay for.

Autonomous content-publishing pipeline that learns from its own audience and rewrites its own prompts (agentic, self-improving)

Agentic Ops Harness | Superpowers Like AI Harness for Ops. Currently works with Hermes Agent

Universal agentic engineering layer for Google Antigravity. Features autonomous multi-agent orchestration, epistemic grounding, 9 strict verification gates, AST blast radius analysis, and target project adaptation.

An agentic AI-powered, markdown-native CRM that turns your Google Workspace activity into a durable, Git-versioned knowledge base where specialized LLM agents compile unstructured signals into structured relationship memory, keeping you fully in the loop as the final decision-maker.

Companion code project for the ebook Vector Database Engineering for AI — Building High-Performance Retrieval Systems for LLMs, RAG, and Agentic AI.

Agent harness to think ideas through and write specs that match what you actually mean.

Agentic RAG system with LangGraph, hybrid BM25+FAISS retrieval, cross-encoder reranking, Corrective RAG, FastAPI, RAGAs evaluation, and Docker deployment

Penalize the Path, Reward the Outcome — verifiable per-action penalties as a dense channel for deployable, sample-efficient agentic RL (GRPO). Paper: arXiv:2607.07435

Ultimate vibe coding plugins for Hermes AI agent — Effect-ts, LSP, Semble code search & Graphify knowledge graph + t/s status bar. 23 tools, 5 plugins, stdlib core.

Developer infrastructure for discovering and comparing LLM models and providers

real-time and file-based speech-to-text desktop app powered by OpenAI Whisper with a modern Flet GUI

GhostPour — Open-source LLM API gateway with auth, multi-provider routing, and subscription-based access control. The invisible hand for your iOS app.

Retroactive Claude Code Skill usage analytics. Single-file Python CLI, reads your existing ~/.claude transcripts, counts real Skill tool_use invocations (not prose mentions). No hook to install ahead of time.

Real-time biofeedback engine for macOS — reads cognitive overwhelm from your face and webcam at 30 FPS, then lets Claude restructure your workspace.