A community‑reconstructed version of Claude Code that restores the original CLI and lets you run Claude‑family models through GitHub‑hosted inference endpoints (GitHub Models and Copilot). Install with Bun, pick a provider/model via `/provider` or command‑line flags, and use the restored tool‑use loop. Some private modules are shimmed, so behavior may differ from the proprietary original.

hoop is an open‑source sidecar that sits between AI agents and their data sources, providing runtime data‑masking, guardrails, optional model‑based risk analysis, and human review escalation—all configurable via a single YAML file and manageable at scale through a lightweight control‑plane UI.

An open-source SDK that provides a full voice stack for AI agents, integrating LLMs, STT, TTS, and telephony carriers into a single API.

A collection of open-source, on-device text-to-speech models that enable natural-sounding speech and instant voice cloning locally on mobile and embedded devices.

An offline Android TTS engine that enables local voice cloning from short audio samples for private, on-device text-to-speech synthesis.

A native Linux speech-to-text tool providing fast, private, and system-wide dictation using local or cloud-based AI models.

GraphRAG‑SDK is a Python SDK that builds a knowledge graph in FalkorDB from your documents and then answers questions by traversing that graph. It adds provenance edges, multi‑hop retrieval, and optional abstention to dramatically cut down LLM hallucinations, and it ships with benchmarks showing it outperforms standard vector‑RAG. Install via pip, run a FalkorDB container, ingest texts (or PDFs/markdown), call `finalize()`, and query with `completion()`. The library supports schemas, incremental updates, custom providers, and is Apache‑2.0 licensed.

Arbigent (Arbiter‑Agent) is an open‑source AI‑agent testing framework for mobile, web and TV apps. It lets QA engineers create test scenarios in a visual UI using natural‑language goals, while developers can run the same scenarios programmatically from YAML files or the CLI. Features include task decomposition, cross‑platform support, model‑choice flexibility, image‑based assertions, reusable scenario libraries, Maestro/YAML integration, and extensibility via a Model Context Protocol. Installation is via a macOS binary for the UI or a Homebrew‑installed CLI (with an optional wrapper script for pinned versions). The tool is free, community‑driven, and targets teams that want high‑fidelity UI testing powered by LLMs, accepting the trade‑offs of slower speed and AI cost.

A terminal-based eBook and document reader that synchronizes text-to-speech playback with word highlighting and auto-scrolling.

SearchCLI is a command‑line client for Volcengine’s AI Search service. It lets you manage datasets, create search apps, and run search, recommendation, and chat‑style retrieval from the terminal. The tool also supports LLM‑based relevance judging and provides “Viking skills” – installable command bundles that external AI agents can invoke, enabling automated data onboarding, configuration, and quality‑verification workflows.

Neuron AI is a Composer‑installable PHP framework for building full‑featured, agentic applications. It provides an `Agent` base class (with memory, toolkits, and multi‑provider LLM support), a `RAG` class for retrieval‑augmented generation, a low‑level `Workflow` engine for custom pipelines, structured‑output typing, MCP connector for remote tools, and optional observability via Inspector. The library targets PHP 8.1+, supports 20+ LLM back‑ends, and includes CLI generators, docs, and Docker‑based test setup.

SwitchAI is an Android app that lets you select, launch, and set a default among dozens of AI digital‑assistant apps (ChatGPT, Claude, Gemini, Alexa, etc.) via a Material 3 UI, home‑screen widgets, and Quick‑Settings tiles.

Intel AI Playground is an open‑source desktop suite that bundles local LLM chat, image/video generation, code assistance, audio transcription, and tool‑calling into a single UI. It runs on Intel Core Ultra CPUs, Intel Arc GPUs, or NVIDIA RTX cards, and can optionally fall back to cloud APIs (ChatGPT, Claude, Gemini, Grok). The beta installer sets up back‑ends like Llama.cpp, OpenVINO, and ComfyUI; models are downloaded on demand via a built‑in Model Manager. Targeted at prosumers and privacy‑focused users who want an offline, all‑in‑one generative‑AI workstation.

A Python package for simulating room acoustics and implementing audio array processing algorithms, used to generate synthetic audio data and test signal processing techniques.

A Python‑based Agent Skill that lets LLM agents interact with Kaggle: fetch competition info, summarize top write‑ups, search discussions, download and reproduce notebooks, submit solutions, and upload datasets—all via natural‑language commands.

Linux-Voice-Assistant is an open-source tool that turns a Linux device into a voice satellite for Home Assistant. It handles wake word detection locally and streams your speech to Home Assistant's Assist, letting you build your own smart speaker with timers, announcements, and conversation support.

A self-hosted server with a Web UI and OpenAI-compatible API for Resemble AI's Chatterbox TTS family, featuring voice cloning, multilingual support, and audiobook generation.

An open-source singing synthesis editor for mobile devices that supports DiffSinger, UTAU, and Vogen voicebanks.

AgentRQ is an open‑source platform that lets humans and LLM agents (Claude, Codex, Antigravity, etc.) share a real‑time task workspace. It provides a web UI and native desktop app, a Go‑based backend exposing a Model Context Protocol (MCP) that agents call to create/update tasks, fire events, read/write notes, and more. Features include a Kanban board, scheduled/recurring tasks, drag‑and‑drop event‑driven workflows, on‑device auto‑title generation, speech‑to‑text, and extensive keyboard shortcuts. Extensions can add custom pages or actions. The repo includes Docker/self‑host instructions, a one‑liner desktop installer, and guides for integrating Claude via `.mcp.json` and for bridging other agents through an ACP gateway.

A JavaScript library for voice activity detection that allows developers to run callbacks on audio segments containing user speech.

Sage is a security plug‑in for AI coding assistants (Claude Code, Cursor, VS Code, OpenClaw, OpenCode). It intercepts tool calls—shell commands, URL fetches, file writes, package installs—and runs multiple threat checks (cloud URL reputation, 300+ local YAML heuristics, a prompt‑injection ML model, supply‑chain analysis, and Windows AMSI). If a request looks unsafe, Sage blocks it and logs the decision. Installation is done via the respective assistant’s marketplace or npm, and the project is open‑source under Apache 2.0.

LatticeDB is an embedded, single‑file property‑graph database that also supports HNSW vector similarity and BM25 full‑text search. It offers a unified Cypher query language, ACID transactions, durable change‑feeds, and bindings for Python, TypeScript, Go, and Java. Benchmarks show sub‑microsecond node lookups, sub‑millisecond 10‑NN searches on 1 M vectors, and microsecond full‑text queries, all without a server. Ideal for local knowledge graphs, agent memory, and RAG pipelines, but not suited for multi‑writer or distributed workloads.

A framework for generating expressive, long-form, multi-speaker conversational audio from text, capable of producing podcasts up to 90 minutes long.

NornicDB is an open‑source Go database that merges Neo4j‑compatible graph traversal, vector‑search (HNSW, GPU‑accelerated), and MVCC‑based historical reads into a single engine. It offers Bolt/Cypher, REST, GraphQL, and Qdrant‑compatible gRPC APIs, supports hardware acceleration (Metal, CUDA, Vulkan), and includes AI‑focused features like decay/retention policies and auto‑relationship generation. Designed for agent memory, Graph‑RAG, and audit‑heavy knowledge stores, it claims dramatically lower latency than Neo4j and provides Docker images for quick deployment.

Writing‑Agent is a Windows‑compatible, Claude‑Code‑driven tool that turns article creation into a multi‑stage, reproducible pipeline with evidence tracking, fact‑checking, and AI‑tone removal. It works via CLI, a plugin, or a preview desktop app and supports low‑cost Anthropic‑compatible models such as deepseek‑v4‑flash.

A high-accuracy speech-to-text system that allows users to choose between verbatim transcription (including fillers) and clean, intended text, featuring precise word-level timings.

A graphical user interface for faster-whisper and whisperX that allows users to transcribe audio and video files into various subtitle formats.

A Wyoming protocol server for faster-whisper that enables local speech-to-text with dynamic name biasing for Home Assistant entities.

A long-form spoken dialogue generation model that transforms scripts into expressive multi-party conversations with support for up to 60 minutes of continuous audio.

CCManager is a terminal UI that lets developers run, monitor, and organise multiple AI coding‑assistant sessions (Claude, Gemini, Codex, etc.) across one or many Git worktrees. It creates and deletes worktrees, can copy ignored files and Claude conversation data, persists session state across restarts, offers visual status indicators, configurable keyboard shortcuts, automation hooks, devcontainer support, and a multi‑project mode. Install with `npm i -g ccmanager` and launch via `ccmanager`.

Vestige is an open‑source, locally‑run memory layer for LLM coding agents. It records decisions as “memories”, merges duplicates, flags contradictions, and can backtrack from a bug to the earlier decision that caused it (backfill). All data stays on the developer’s machine (SQLite), with optional encrypted sync via a paid Pro tier. Installation is a single npm command; the tool ships as a 25 MB Rust binary that downloads an embedder and reranker once, then works offline. Benchmarks show Vestige helping agents converge on correct fixes far more reliably than plain vector‑search RAG. Licensed under AGPL‑3.0.

AI DevKit is a local npm‑based control plane that unifies configuration, monitoring, cross‑agent messaging, searchable SQLite memory, and composable engineering “skills” (requirements, design, testing, review, etc.) for many AI coding assistants such as Claude Code, Gemini CLI, Codex CLI, Cursor, and GitHub Copilot. It lets teams manage multiple agents from a single console, keep project conventions in a searchable store, and enforce a senior‑engineer workflow without modifying the underlying LLMs.

A multi-voice, prompt-controlled text-to-speech engine that enables emotional synthesis and voice cloning for English and Chinese.

Alphastream is an AI‑driven financial analysis web app. It uses DeepSeek/any OpenAI‑compatible LLM with LangGraph to run 14 specialized agents (technical, fundamental, sentiment, risk, etc.) plus four investor‑persona agents. Real‑time market data comes from dual adapters (AKShare & BaoStock) covering Chinese, US, Hong‑Kong, crypto and macro sources. The Next.js + React front‑end shows a dark glass‑morphism UI with chat, dashboards and chart cards, while the Flask backend streams agent progress via SSE. All services can be launched with a single Docker‑Compose command.

Dapr Agents is a Python SDK built on the Dapr runtime that lets developers create scalable, fault‑tolerant AI agents. It provides durable workflows, actor‑based execution, built‑in data source bindings, multi‑agent communication, and automatic tool discovery, all with Kubernetes‑native deployment and strong security.

Oumi is an open‑source Python platform that covers the entire lifecycle of foundation models—data prep, training (SFT, LoRA, QLoRA, GRPO), evaluation, and deployment—via a unified CLI and a library of ready‑made YAML recipes for many text and vision‑language models.

Synalinks is an open‑source Python framework that lets you compose, train, and deploy neuro‑symbolic language‑model applications (RAG, agents, recursive pipelines) using a Keras‑style declarative API. It provides sandboxed agents, embedded graph/SQL knowledge bases, in‑context reinforcement‑learning optimisation of prompts, and easy model switching via LiteLLM. The library supports FastAPI/FastMCP deployment, MLflow observability, and hyper‑parameter search, targeting AI developers, researchers, data scientists, and learners.

A collection of ComfyUI nodes for separating audio into stems using Hybrid Demucs and performing audio editing tasks like tempo matching and cropping.

DSH Desktop is a community‑maintained desktop distribution that packages the official DeepSeek Harness AI runtime, a built‑in plugin marketplace, a better sidebar workbench (file tree, editor, Git, terminal), task board, and optional mobile/remote access. It runs the DSH web UI inside an Electron host, uses a single profile for all data, and ships 16 verified plugins out‑of‑the‑box. Installers are available for macOS Apple Silicon and Windows x64; other platforms can be built from source.

A modular Swift SDK for audio processing using MLX on Apple Silicon, providing easy integration of TTS, STT, VAD, and audio codecs into macOS and iOS apps.

A Go‑based workshop that incrementally builds a Claude‑powered coding assistant, adding tools for file reading, directory listing, shell execution, file editing, and code search.

An open-source voice dictation tool that converts speech to text anywhere on your system, supporting local transcription and LLM post-processing for privacy and speed.

A private, cross-platform AI voice transcription and note-taking app that uses OpenAI Whisper to provide offline speech-to-text capabilities.

TorchGeo is a PyTorch library that makes working with satellite and aerial imagery easy. It offers ready‑made geospatial datasets, smart samplers for patch extraction, multi‑spectral pre‑trained models, and tight integration with PyTorch Lightning, letting researchers train and evaluate remote‑sensing models with just a few lines of code.

A benchmarking suite for local text-to-speech models that evaluates speed, objective quality scores, and human-preference rankings across various hardware rigs.

Commerce Agent Bench is an open‑source benchmark for evaluating LLM‑based agents on 107 realistic, stateful commerce tasks (CLI, browser, API, file). It provides Docker‑isolated mock services, a reproducible OpenClaw harness, multiple provider routes, and a public leaderboard. The repo includes installation instructions, example runs, and a reproducibility contract, but does not ship any agents itself.

EasyAIoT is an open‑source cloud‑edge‑device platform that unifies AI video analytics, IoT device management, and enterprise integration. It runs on tiny edge boxes (≈1 GB RAM), supports both professional IP cameras and consumer P2P cameras (Tapo, Tuya, Ring, etc.) via a built‑in RTC module, provides native YOLO/SAM inference, federated edge scheduling, and a one‑click installer (PANEL) for on‑site acceptance. Installable on Linux, Windows, macOS, and ARM, it offers web, mobile, and SCADA‑style dashboards for real‑time monitoring and alerting.

A React component library built on shadcn/ui that provides pre-built UI elements like waveforms and orbs for building audio and AI agent applications.