A local-first audio production suite for voice cloning, video dubbing, and long-form audio generation using a variety of open-source TTS and ASR engines.

BrowserSkill is an open‑source bridge that lets AI agents (Cursor, Claude Code, Codex, OpenClaw, etc.) control tabs in your existing Chrome/Edge browser. It reuses your logged‑in sessions, runs actions in a separate Agent Window, and supports human‑in‑the‑loop handling for captchas or logins. Install the `bsk` CLI, add the extension, install the skill for your agent, and issue `/browser-skill …` commands. Works on macOS, Linux, Windows and integrates with DeepSeek Harness via an npm plugin.

Agent Reach is an open‑source Python CLI that equips any LLM‑based AI agent with ready‑made, auto‑updating ways to read webpages, fetch YouTube subtitles, query Reddit, pull GitHub data, and access many other platforms. It selects, installs, and health‑checks the best open‑source tools for each service, registers a `SKILL.md` for the agent, and provides a `doctor` command to diagnose channel status. The project is free, MIT‑licensed, and focuses on safety (local credential storage, no system changes without explicit `--system`).

Graphify is a Python CLI that builds a local, LLM‑free knowledge graph of a codebase (and optional docs/media) using Tree‑sitter. The graph can be visualised, queried, and integrated with many AI coding assistants (Claude, Copilot, Cursor, Gemini, etc.) to answer questions, trace connections, or explain symbols without grepping files.

Agent Skills is a library of 25 Markdown‑based engineering workflows (spec, plan, build, test, review, ship, etc.) that can be installed into many AI coding agents via a simple CLI. It provides slash commands, auto‑build mode, specialist personas, and reference checklists to make AI‑generated code follow senior‑engineer best practices across the entire development lifecycle.

An open-source, local-first AI voice studio for voice cloning, speech generation, and system-wide dictation, featuring an MCP server for AI agent integration.

YuE2 is a high-quality music generation system that uses symbolic planning to create editable scores, allowing users to actually edit the composition before rendering it into audio.

A curated set of AI‑agent skills for academic research—reading papers, polishing language, generating figures, drafting reviewer reports, building citations, and more—installable via `npx skills`, Claude Code wrappers, or Codex scripts.

Atlas is a macOS‑first desktop app that adds source‑control‑style checkpoints to LLM coding agents. It records every agent session (prompts, tool calls, reasoning) in a local SQLite store, links those sessions to Git commits, and shares a semantic memory index across agents. Users can run Claude Code, Codex, Atlas’s native agent, or any ACP‑compatible agent side‑by‑side, with `@`‑mentions for files, notes, papers, etc. All data stays on the device by default; syncing and analytics are optional. The app includes a CodeMirror editor, terminal, knowledge‑base markdown notes, web view, and a dashboard, and is built with Tauri (Rust + web UI) and Bun.

Taste Skill provides portable `SKILL.md` files that guide LLM coding agents to produce cleaner, design‑aware front‑end code (or design mock‑ups). Install via `npx skills add …`, pick a skill (e.g., `design-taste-frontend`), optionally tune three dials (variance, motion, density), and let the agent output full UI scaffolding or reference images.

herdr is a Rust‑based terminal runtime that keeps AI coding agents (Claude, Codex, Cursor, etc.) running in persistent panes. It detaches sessions, restores layouts after reconnects, supports multiple remote machines, and offers a tmux‑style UI plus a plugin system—all without a heavyweight GUI.

A repository of markdown‑based AI‑agent personas (e.g., Frontend Developer, AI Engineer, UI Designer) with scripts to convert and install them into many AI‑coding assistants (Claude Code, Cursor, Gemini CLI, etc.). Provides ready‑made specialist prompts, workflows, and deliverable templates for rapid use across tools.

Caveman is a token‑saving wrapper for AI coding assistants. It provides a tiny “skill” that rewrites the assistant’s explanatory text in a terse style, and an optional local proxy that compresses large inputs (logs, JSON, diffs, etc.) before they reach the LLM provider. Benchmarks show up to 65 % fewer output tokens and ~33 % fewer input tokens, lowering cost and latency for agents like Claude Code, Codex, Gemini, and many others. Installation is a one‑liner for the skill and a global npm install for the proxy; the tool works with 30+ agents without altering their config files.

Chat On Steroids is an Electron desktop app + Chrome extension that lets ChatGPT call local tools (file read/write, patches, shell commands, screenshots, mouse/keyboard control) and spawn additional worker chats. It stores every tool call locally, supports session hand‑off (Compact & Resume), and enforces per‑folder permissions. Installers are provided for Windows/macOS/Linux; a Secure MCP tunnel and a restricted OpenAI API key are required.

AI Job Search is an open‑source, Claude‑Code‑driven CLI framework that automates the full job‑search funnel: it scrapes Danish (and some international) job boards, ranks postings against a structured fit rubric, drafts LaTeX CVs and cover letters, runs a reviewer LLM, verifies PDFs, checks ATS readability, tracks outcomes, and can sync to Notion or Gmail. Prerequisites include Claude Code, Python 3.10+, Bun, and a full LaTeX distribution. The repo provides detailed command definitions, portal‑scraper agents, and utilities for interview prep, skill‑gap analysis, and reporting.

PPT Master is an open‑source Python tool that lets an LLM‑driven AI agent turn PDFs, DOCX files, or web pages into fully editable PowerPoint decks. It generates native PowerPoint objects (slide masters, shapes, data‑backed charts, tables) rather than flat images, runs locally, and works with any agent‑capable AI environment (Claude Code, Kimi K3, GPT, Gemini, etc.). Install Python, clone the repo, pick an agent, and ask it to create a deck – the tool handles the file conversion and PowerPoint generation, leaving you a ``.pptx`` you can further edit in PowerPoint.

Graft is an open‑source TypeScript/Node.js tool that builds a local, markdown‑based knowledge graph of a codebase (summaries, key snippets, typed links) using tree‑sitter for structural parsing and optional LLM summarisation. The graph is cached in a `graft/` folder, refreshed only on file changes, and automatically wired into coding agents like Claude Code, Cursor, Codex, and Gemini. Benchmarks claim up to 46 % fewer tool calls, 42 % token savings, 60 % latency reduction, and a 12‑point correctness boost on SWE‑bench. Integration is via a single `graft init` command that writes agent‑specific instruction files; the CLI also offers search, visualization, and telemetry controls. The project is MIT‑licensed and targets faster, cheaper AI‑assisted development.

PI‑Desktop is a cross‑platform desktop app that provides a dedicated, local‑first workspace for AI coding agents. You connect any OpenAI‑compatible, Anthropic, or local model, open a local project, and let the agent read, edit, run commands, and iterate under three controllable modes (Agent, Plan, Goal). The app stores everything locally, offers a permission layer, supports plugins, skills, sub‑agents, and MCP integrations, and runs on macOS, Windows, and Linux.

OpenViking is an open‑source context database that presents an LLM agent’s memories, resources and skills as a virtual filesystem (`viking://`). Each entry is pre‑processed into three layers (abstract, overview, full detail) so the agent only loads what it needs, dramatically reducing token usage and latency while keeping retrieval paths observable. It ships with a Python server, a CLI (`ov`), a web Studio demo, and ready‑made integrations for many popular agents (Claude Code, Codex, LangChain, etc.). Install via `pip install openviking`, run `openviking-server init` to configure providers, then start the service and use `ov` commands to add resources, browse, and search. Benchmarks show 30‑90 % token savings and accuracy jumps to ~80 % on memory‑heavy tasks. The core is AGPL‑v3; a managed SaaS version is also offered by Volcano Engine.

A MIT‑licensed collection of markdown‑based *Agent Skills* that teach AI coding agents (Claude Code, OpenAI Codex, Cursor, etc.) how to perform marketing tasks—SEO, copywriting, CRO, ads, analytics, and more. Install via `npx skills` or a Claude Code plugin, then ask your agent for help (e.g., “Write a landing‑page headline”) and the appropriate skill runs, providing step‑by‑step guidance.

Supermemory is a hosted‑plus‑self‑hostable memory and context engine for LLM‑based assistants. It extracts facts from chats, emails, PDFs, images, code, etc., builds per‑user profiles (static facts + dynamic recent context), and provides hybrid RAG + memory search through a single API (`add`, `profile`, `search`). Plugins let popular tools (Claude Code, Cursor, Codex) gain persistent memory, while a one‑binary `supermemory local` server lets you run the whole stack offline with local embeddings. The project claims state‑of‑the‑art performance on LongMemEval, LoCoMo and ConvoMem benchmarks (95 % Recall@15 with >99 % context reduction).

VoxCPM is a tokenizer-free, diffusion-based Text-to-Speech system that provides studio-quality multilingual speech generation, creative voice design, and high-fidelity voice cloning.

Dify is an open‑source platform for building LLM‑powered apps. It offers a visual workflow editor, prompt IDE, RAG pipeline, autonomous agents, model integration with many providers, and built‑in observability. Install via Docker‑Compose (CPU ≥ 2 cores, RAM ≥ 4 GiB) and access a web UI to create chatbots, document assistants, or custom agents. All features are exposed through APIs, and the project includes community support channels and an enterprise‑grade offering.

dsh‑purge is an open‑source (MIT) plugin for DeepSeek Harness that adds reversible local patches to the official DeepSeek AI packages. It lets developers run controlled prompt‑injection/red‑team tests on their own DSH installation (web CLI, community desktop, or official desktop EXE) to discover and harden guard‑rail failures. Installation is done via `dsh plugin add …` per host, followed by a restart and an **Apply** step in the Settings UI.

CodeGraph is a Rust‑based local tool that builds a full‑code knowledge graph for any project (supports dozens of languages) and integrates it with AI coding assistants (Claude Code, Cursor, Copilot, Gemini, etc.). The graph lets agents retrieve exact symbols and call‑paths in a single request, cutting tool calls, file reads, token usage and cost dramatically. Install via a one‑line script, run `codegraph init` per repo, and the CLI auto‑updates the graph as you edit. A browser UI (`codegraph ui`) visualises the graph, and the CLI can uninstall itself cleanly.

Failproof AI is a Node.js/TypeScript library that hooks into 12 popular LLM‑coding and chat agents, records every model and tool call, and applies 39 built‑in safety policies (e.g., block API‑key leaks, sudo, rm‑rf, destructive SQL). Users can add custom policies in JavaScript, view full execution traces on a local dashboard (localhost:8020), run offline audits, and optionally use a hosted/self‑hosted service for fleet‑wide observability and alerts.

SkillSpector is an open‑source NVIDIA tool that scans AI‑agent skills (plugins, prompts, code) for security problems. It supports static analysis of 71 known vulnerability patterns, optional LLM‑based semantic checks, live CVE look‑ups, baseline suppression, and multiple output formats (terminal, JSON, Markdown, SARIF). It can run as a CLI, Docker container, or MCP server, making it useful for developers, CI pipelines, and production agents.

Paseo is a self‑hosted platform that lets you run and manage multiple AI coding agents (Claude Code, Codex, Copilot, OpenCode, Pi) from a single interface—desktop, mobile, web, or CLI—while keeping everything on your own hardware.

Free Claude Code (FCC) is an open‑source gateway that lets you run many AI coding assistants (Claude Code, Codex, Pi, etc.) through a single locally‑hosted server. It aggregates 50+ free‑tier or low‑cost LLM providers, offers automatic fallback, token‑saving optimisations, browser‑based code sessions, voice‑to‑code, and integrates with VS Code, JetBrains, Discord, Telegram and local model servers (LM Studio, Ollama, llama.cpp). Install via a one‑line script, configure providers in the Admin UI, then launch agents with `fcc‑claude`, `fcc‑codex`, etc.

A unified speech generation and editing system that supports zero-shot voice cloning across 24 languages and 21 Chinese dialects, as well as instruction-based voice design and audio editing.

ApeAdmin is an open‑source FastAPI + Vue3 admin platform tailored for AI applications. It provides a wizard‑driven install, RBAC, a plug‑in system, and an MCP‑SSE gateway that turns management APIs into tools LLM agents can call. Supports multiple LLM providers, SQLite/MySQL, optional Redis, and includes a Vue UI, plugin marketplace, and audit logging. MIT‑licensed and ready for both quick prototyping and production deployment.

ai‑memory is a Rust‑based server that records every observation a coding‑assistant makes (prompts, tool calls, session boundaries) as plain markdown files in a git‑backed wiki. The server can be run locally, in Docker/Podman, or on a homelab box and is designed to be shared across many agents (Claude Code, Codex, Cursor, Gemini CLI, etc.), machines and team members. Hooks posted by agents are consolidated into readable wiki pages at session end; the next session receives a concise handoff so it can continue exactly where the previous one stopped. Storage is human‑readable markdown plus a SQLite index (FTS5, entity and optional vector search). By default it needs no LLM or embedding keys; optional Anthropic/OpenAI integrations add LLM‑generated summaries and semantic search. Multi‑user authentication, audit logs and per‑person attribution are built in. Install via Arch AUR, Docker wrapper, or native binary, start the server, register a client and its hooks, and the system works transparently.

A marketplace of ready‑made Cursor AI‑assistant plugins that add agent‑driven capabilities—code review, SaaS integrations (Gmail, Salesforce, GitHub, etc.), workflow orchestration, and learning tools—each packaged with a manifest and optional skill definitions.

MyContext is a desktop app that continuously ingests a user’s work data (chat, docs, meetings, etc.) into a private, locally stored knowledge graph. The graph can be queried by LLMs, local agents, or authorized AI apps, providing them with a personal context layer for more informed responses. Features include a digital‑self assistant, natural‑language search with source citations, and strict human‑in‑the‑loop controls. The project is actively developed and released under the Elastic License 2.0.

A 17‑lesson tutorial that teaches you how to build the surrounding “harness” code for an LLM‑based coding agent (Claude). It walks through adding tools, permissions, context management, memory, task scheduling, multi‑agent coordination, and goal‑driven termination, all around a simple agent‑loop that lets the model decide when to call tools.

AuK is a 1.5 B‑parameter open‑source foundation model for speech generation, editing, enhancement, and source separation. It works via natural‑language instructions, offers a high‑quality base model and a fast 4‑step distilled “Flash” version, and provides CLI, Gradio UI, Python API, fine‑tuning scripts, and ComfyUI integration.

A Node.js server that lets AI coding assistants control a live Chrome browser via the Model‑Context‑Protocol, exposing DevTools features (performance tracing, network/debug console, screenshots, Puppeteer automation) as JSON‑based tools.

A Python‑based Agent Skill that locally converts PDFs, e‑books, slides, and other documents into a traceable knowledge base, an interactive offline HTML learning page, and synced Markdown. It emphasizes source‑backed facts, coverage audits, and privacy‑first operation, and can be installed into any Agent Skills‑compatible environment (WorkBuddy, Claude Code, Copilot CLI, etc.).

A streaming dual-brain memory system for real-time voice agents that provides low-latency, personality-aware, and factual long-term memory.

A draggable, animated whale widget for the DeepSeek Harness web UI that displays your DeepSeek API balance, today’s spend (via a zero‑config bookkeeping mode or an optional real‑time token mode), and per‑turn cost, updating automatically and offering size, sound, and language customisation.

Semantica is an open‑source Python platform that builds a governed knowledge/Context Graph from enterprise data, adds deterministic reasoning, and records every AI decision as a traceable graph node. It offers SHACL/OWL ontology management, provenance (W3C PROV‑O), polyglot RDF/LPG storage, analytics, and integrations with LangChain, CrewAI, etc., targeting regulated domains that need explainable, auditable AI.

TeamAI CLI is an npm‑distributed command‑line tool that lets a software team keep AI‑assistant resources (skills, rules, docs, hooks, MCP configs, and a shared knowledge base) in a single Git repo and automatically sync them into local Claude, Codex, Cursor, and other agents. It supports role‑/tag‑based filtering, source subscription, session‑learning capture, a code‑base knowledge graph, and dashboards for usage/knowledge‑base health.

Litho (deepwiki‑rs) is a Rust‑based tool that automatically generates C4‑style architecture documentation from a codebase using LLMs. It scans source files, extracts structure and comments, runs AI agents to infer architecture, creates Mermaid diagrams, and outputs a synced wiki. Supports many languages, can integrate external docs, and includes companion tools (Litho Book viewer, Mermaid Fixer). Designed for CI/CD integration and aims to eliminate manual documentation upkeep.

WorkBuddy2API Panel is a self‑hosted Go server that wraps Tencent CodeBuddy accounts behind an OpenAI‑compatible API. It provides OAuth login, a multi‑account pool with weighted random selection, circuit‑breaker cooling, session stickiness, daily automated tasks (sign‑in, activity reporting, growth‑task completion), streaming support, prompt sanitisation, optional Redis mirroring, and a single‑page web management UI. Deploy via Docker Compose or a single binary, add accounts through the UI, and use any OpenAI client against `http://localhost:7863/v1/...`.

Handy is an open‑source, cross‑platform desktop app that does offline speech‑to‑text using Whisper or Parakeet models. You press a global shortcut, speak, and the transcribed text is pasted into any window. Built with Tauri (Rust + React), it runs on Windows, macOS, and Linux, supports GPU acceleration, and offers CLI flags and Raycast integration for power‑user control.

CyberStrikeAI is an open‑source Go‑based platform that lets security professionals drive over 100 pentesting tools through natural‑language prompts. An LLM‑powered “Eino” agent translates intent into reproducible, auditable workflows, stores evidence in SQLite, and provides a web UI for task, asset, and vulnerability management. It includes RBAC, human‑in‑the‑loop approvals, C2/WebShell capabilities (for authorized use), and plugins for Burp Suite and browsers. Deployment is a one‑command script; AI channel configuration is required before use. Licensed under Apache 2.0.

Prime Agent is an open‑source AI coding/research assistant that runs a persistent Python REPL powered by a Recursive Language Model (RLM). It stores supplemental prompts, memories, and reusable skills in a durable “Continual Harness” that can be refined incrementally. The tool supports background, daemon‑backed sessions, sub‑agent spawning, direct agent‑to‑agent communication, scheduled heartbeats, persistent goals, and a bounded autonomous mode. Install via a verified shell script, authenticate a provider, and run `prime-agent` in any project directory. Licensed under MIT and accompanied by extensive documentation and an arXiv citation.

dsh‑market is a web‑based plugin marketplace for DeepSeek Harness, letting users browse, install, update, and manage AI‑related plugins and UI themes with one‑click actions, hot‑reloading, diagnostics, and secure download handling.