Custom agents

Open CoChat is an open-source multiplayer real-time voice chat platform with integrated AI assistant, supporting voice/text dual-mode interaction where AI can intervene in group chats in real-time, generate and edit files. Open CoChat是一个开源的多人实时语音聊天平台，集成 AI 智能助手，支持语音/文字双模式交互，AI可实时介入群聊、生成和编辑文件

🦞 One command to stop your Mac from sleeping — so OpenClaw never loses connection. 一条命令让 Mac 永不休眠，OpenClaw 远程操控不断线。

🤖 D.A.V.I.D AI - Advanced AI assistant with voice control, gesture recognition & multi-language support. Privacy-first, offline-first, 15 languages. By Nexuzy Tech Ltd.

Python runtime checks for AI agents: budgets, loops, retries, and local traces. Zero runtime dependencies. MIT.

AI-powered PDF Q&A — RAPTOR indexing, Llama 3.3 70B, hybrid BM25+FAISS retrieval

A small Telegram bot powered by Gemini AI to retrieve informations from a custom set of documents

RushDB Python SDK

♊ Voice-Interactive AI Digital Twin powered by Google Gemini 3. Talks to recruiters via Neural TTS, visualizes architecture, and defends against prompt injection.

RAG pipeline for SEC EDGAR filings (10-K, 10-Q) with XBRL parsing, pgvector semantic search, and cross-encoder reranking

Governed knowledge store for autonomous organizations. Trust tiers, content-addressed storage, Merkle verification, knowledge lifecycle, air-gap native.

VIC-style tiered retrieval demo with native KIOXIA AiSAQ for low-DRAM, storage-aware semantic search.

企业AI开源应用套装(LLM + Dify/n8n + MCP Server + OAuth2)

Complete Workspace Template for OpenClaw - Full agent lifecycle with unified memory system (Markdown + SQLite), self-evolution, RAG. Not for SubAgent/Skill use.

A Multi‑Hop Question‑Answering framework for Retrieval‑Augmented Generation (RAG) using Triple Placeholder Resolution and Entity‑Centric Summaries.

Jenkins MCP server in Go — expose Jenkins jobs, builds, and console logs to LLMs.

Type-safe TypeScript SDKs for Meta (Facebook, Instagram, Threads), TikTok, and Google Ads — fetch-based, auto-generated from official specs, AI-agent ready. Powering openpromo.app.

MCP drawing application with TypeScript server and React frontend for real-time canvas visualization through Claude Desktop.

Machine-readable merchant verification infrastructure for AI shopping agents and agentic commerce systems.

A headless Python CLI for interacting with the Tidal music streaming API, designed for LLM agent automation. Supports OAuth authentication, artist/album/track search, playlist management (create, rename, delete, add/remove tracks), and library favorites. Outputs human-readable text or structured JSON via a --json flag.

Persistent code graph for AI agents. Tree-sitter → SQLite → MCP. Agents store understanding once, skip re-reads forever. 51–1500x fewer tokens.

Paper Implementation: Semantic Tool Discovery for Large Language Models: A Vector-Based Approach to MCP Tool Selection https://arxiv.org/html/2603.20313v1

A collaborative BrAPI v2.1 workspace for multi-agent research via MCP. Search studies, germplasm, genotypes, and more across Breedbase, T3, Sweetpotatobase, and any BrAPI v2-compliant server.

Operating mechanisms for Claude Code. Not rules. Not principles. Mechanisms with triggers, retry logic, and failure notifications. 20+ mechanisms, MIT.

Hammy is a code base mapping tool that lets agents talk to your code in a way they understand with an AST, call graph, and VCS history.

Cross-conversation memory for Claude Code and Cowork

Side-by-side A/B testing for CLAUDE.md, skills, and agents, with live transcripts and AI-judged results.

MCP Excel Server — A Python-based Model Context Protocol (MCP) server using openpyxl to enable LLMs and agents to interact with Excel files: create workbooks, read/write sheets and cell ranges, apply formulas, take screenshots (Windows), and more. Features fully automatic tool registration via decorators and seamless integration with AI workflows

A modern, real-time quota monitor for MiniMax API. Features a semi-transparent floating UI, precise RPM tracking, smooth animations, and persistent configuration.

marketo-cli: Marketo CLI for the Marketo REST API. The installed command is mrkto.

Mechanistic interpretability workflow platform for activation capture, analysis, interventions, and reports.

macOS developer tool that snapshots your project into one readable file and applies AI-edited changes back. Deterministic, LLM-friendly dump/undump.

Semantic code context for AI agents.

Workflow orchestration for AI coding assistants. Creates isolated git worktrees & tmux sessions for Claude Code. Practical infrastructure for AI-enhanced development.

AI-powered GitHub Issue triage Action — auto-classify, label, detect duplicates, and reply

Vibe-coded experiment in LLM-driven character simulation. First-class chat AND image generation, world map, soul system, group chats, multi-provider LLM routing, ComfyUI/A1111/cloud backends. Alpha, single-tester. Source-available, non-commercial.

Prompts, skills, and agents that survive contact with real workflows. No vendor loyalty. Occasionally heretical. 🧙🏻‍♀️

Self-hosted AI assistant with web interface, integrated autonomous agents and sandboxed execution

Keep agent CLIs running 7×24 — supervise Claude Code / aider / any prompt-arg agent with 11-layer defenses and structured observability.

Reusable Agent Skill for architecture-as-code with LikeC4: create/update C4-style models, validate, and export diagrams.

Protocol for agentic social capital

Polymarket research MCP server with 10 tools (odds, orderbook, spread, history) + autonomous trading bot framework. Claude Sonnet decision-making, pluggable strategies, SQLite persistence

Fully local AI assistant & coding agent. LM Studio + ChromaDB + Tauri desktop app. Second brain, autonomous bots, emotional drives. Zero cloud, zero telemetry.

Multi-agent OS on Claude Code. Each agent gets a Discord channel, a markdown vault for memory, and launchd for cron. Built for solo founders shipping like a team.

Polywave Agent Skill for Claude Code: parallel AI agents with disjoint file ownership, worktree isolation, and human-reviewed plans. Requires polywave-tools CLI.

The core belief layer for AI agents. MnemeBrain Lite stores structured beliefs instead of text memories — enabling contradiction detection, belief revision, and explainable reasoning.

VELA = Versioned Evidence Lifecycle Architecture: portable Codex research workflow environment with evidence lifecycle, public skills, validators, and HELM-readable state.

Opinionated management layer for teams of OpenClaw/Hermes agents

The deliberation & governance layer for autonomous organizations. Built like an institution: verifiable audit log, Sybil-resistant reputation, delegated identity, and metered access over MPP + x402/ATXP.

Visual team-builder, plan-as-artifact, and live execution graph for autonomous Claude Code agent crews. Spawn a 3-role team, generate a DAG plan, run for hours, watch it ship in your browser.

Zero-infrastructure multi-agent orchestration via tmux + filesystem. 11 CLI backends, one MCP server, zero dependencies beyond the MCP SDK.

10-agent GTM AI team on OpenClaw, sharing a Cognee+FalkorDB knowledge graph, deployed to an Islo long-running sandbox, with a Mission Control HTTP UI

Multi-agent AI system for turning natural language or reference software into Verilog RTL, testbenches, docs, and FPGA onboarding assets.

Agentic customer support system using .NET 10 + Microsoft Agent Framework (MAF). Features 6-agent orchestration (Planning + Specialists + Response), deterministic reasoning pipeline, HITL approval gates, compound query decomposition, and full observability via reasoning traces.

Deploy and run Cloudflare Workers from one orchestrator with dynamic bundling, KV cache, and path-based routing for fast delivery

睿威智测 AI 工程测绘多智能体CLI系统 — 基于 Railwise 深度定制，面向工程测量、结构监测、地铁监测

Persistent memory for AI agents — powered by VectLite. Hybrid search, context compaction, multi-agent swarm, memory teleportation, MCP server, skill engine, decay & conflict resolution, and worktree-aware branch-scoped memory.

Durable execution where agents, services, and humans coordinate as equals. Open protocol + managed cloud.

Markdown-native multi-agent workspace framework. Define agents, skills, and workflows as plain .md files — no runtime required.

CyBroLog: executable compact accountability language for multi-agent coordination

[ICML'26] AgentWebBench: Benchmarking Multi-Agent Coordination in Agentic Web

Secure, voice-first multi-agent AI assistant — Kotlin/Compose Android + FastAPI/LangGraph backend with GraphRAG (Neo4j + ChromaDB), LiveKit WebRTC voice, and Docker Compose infrastructure.

Multi-agent orchestration server for Claude Code with fleet coordination, swarm intelligence, and real-time collaboration

Autonomous AI assistant plugin for Claude Code — 5 MCP servers, 21 skills, 3-tier persistent memory, knowledge graph, multi-agent orchestration with Discord/Telegram integration for 24/7 Everyday working

input: video --> output: subtitles -- AI-powered video translation and terminology management pipeline. Features multi-agent translation, RTL support, and a built-in web UI for consistent subtitles.

An implementation of the paper: "Agent Bazaar: Enabling Economic Alignment in Multi-Agent Marketplaces" with Swarms Primitives

Shared code intelligence for agent fleets — AST-aware semantic search + multi-agent memory with git-concurrent coordination

Self-Adaptive Generation Engine — Agent Development Kit with cognitive routing, evolutionary topology, formal verification, and 4-tier memory

Claude Code Agent Teams Dashboard - Monitor and visualize multi-agent team communication logs

Sprint autopilot for BMad Method — drives stories from plan to PR-ready code with git worktrees, commits, and stacked PRs.

🛠️ Boost your workflow with Vibe Kanban, the tool that enhances productivity by optimizing interaction with coding agents like Claude Code and Codex.

🧬 Evolutionary taxonomy and self-improvement framework for AI agents — Lamarckian inheritance, skills as genes, GENOME.md

Secure multi-agent AI framework in Rust — WASM sandbox, 50+ skills, 14 LLM providers, agent intelligence, compliance modules

AEGIS — Agentic AI Learning Platform with Epistemic State Modeling, Ebbinghaus Forgetting Curve, and Cognitive DNA Adaptation.

Builds an autonomous AI robot with vision, voice, and decision-making capabilities using Python, PyTorch, and CUDA technology.

MCP server for the X (Twitter) API — with built-in safety rails for autonomous operation.

Autonomous AI agent skill for aggressive lead generation and growth hacking.

Powerful Python agent runtime with tools, MCP, Hooks, Skills, Rag, memory, sessions, reasoning, and streaming packets.

Agents should execute whenever possible — open registry of reusable AI agent skills and capability definitions

An open framework for building self-improving AI agent teams. Includes a meta-agent architecture with independent auditing, git-backed evolution, and structured feedback loops. Ships with a complete marketing team, a reusable team-builder skill, and a portable prompt that works with any LLM.

Ouroboros: Agent first software engineering infrastructure. Autonomous AI agents plan, implement, test, review, and merge pull requests through typed contracts and constrained workflows. The system manages and improves itself.

Bopo is a local-first orchestration platform for running AI employees in a company model.

Standalone Playwright CLI for AI agent browser automation - control real Chromium with simple commands, token-optimized screenshots, and reliable interactions

Interactive web presentation and self-guided workshop on AI agents, LLMs and AI-assisted coding

Ship Lovable apps from your terminal — agentic CLI with official API, MCP, and browser backends. No browser needed when you have an API key.

Structured task management using JJ (Jujutsu). Uses empty revisions as TODO markers with [task:*] flags, forming a DAG of plannable and executable tasks.

Temporal reasoning for AI agents — deadlines, schedules, sequences, duration estimation (PERT), and decay models. One .atime file. Rust core + MCP server.

An undetectable multi-LLM companion. Designed for your most critical live calls and assessments.

Search, explore, and export your Claude.ai conversation history. Features AI-powered search, analytics dashboard, multiple export formats, and beautiful web interface.

Made for BA, by BA, with BA

Health Query Chatbot using LLM (Mistral-7B-Instruct)

⚡ OpenDesktop — AI-powered desktop agent that controls your computer. Voice, vision, automation, memory, 50+ AI providers. Open source.

A modern, minimal desktop IDE built with Electron, React, and TypeScript.

Own Your Intelligence. Open-source framework for building AI-automated businesses.

Deploy your own personal OpenClaw on AWS Bedrock AgentCore — serverless, ~$9/mo, one-click CloudFormation, Discord/WhatsApp/Telegram 🦞

Self-hosted web relay to use your local GitHub Copilot CLI session from any browser — phone, tablet, or second computer. Streams tool activity, supports ask_user question cards, file uploads, image attachments, model/mode selection, conversation history, and optional SSH tunnel for remote access.