Explore the directory

A Neovim plugin that adds commands for chatting with, translating with and coding alongside large language models such as GPT, Copilot, ChatGLM, Kimi, DeepSeek, Gemini, Qwen or a local Ollama model, inside the editor.

• Editors

An agent memory management layer, positioned as a high-performance drop-in replacement for Mem0, adding semantic deduplication, long-term dialogue memory, per-character memory profiles and multimodal video reasoning.

• Agent memory

An on-device speech-to-text engine SDK from Picovoice that runs entirely locally across desktop, mobile and web platforms, with bindings for Python, C, iOS, Android, Node.js, Flutter, React Native, Java, .NET and the browser.

• Speech

A web interface, CLI and modular architecture for retrieval-augmented generation, document splitting, web scraping and query workflows, built with Haystack, Ollama, Hugging Face and Elasticsearch, and runnable as a proxy in front of an Ollama instance.

• RAG

A native .NET inference engine for GGUF models that runs autoregressive and text-diffusion large language models, image generation and editing, and video generation, with a console app, browser chat UI and OpenAI/Ollama-compatible HTTP APIs.

• Inference
• Generative media

An open-source memory and observability layer for AI agents, adding persistent memory, automatic loop detection, a hash-chained audit trail and a live dashboard on top of a plain Python agent, LangChain, CrewAI, AutoGen or the OpenAI Agents SDK.

• Agent memory
• Monitoring

An open-source vLLM plugin from Snowflake that combines advanced parallelism, speculative decoding and model optimizations to raise inference throughput and lower latency for large language models and embeddings.

• Inference

A unified MCP runtime that aggregates multiple MCP servers behind one server process, with a CLI mode that lets agents discover and run tools progressively instead of loading every schema up front.

• MCP servers

Robust-U1 is the official implementation of a unified multimodal language model that detects and repairs corrupted visual input before reasoning about it, aimed at robust visual understanding under real-world image degradation.

• Models

A curated list of coding agents, assistants, IDEs and editors, chat and RAG tools, mobile apps, automation, notebooks, terminal tools, model evaluation tools, package managers and libraries that work with Ollama.

• Curated lists

An MCP server that exposes DrissionPage's Chromium/CDP browser automation as 69 typed, atomic tools for Claude Code, Codex and other MCP clients, including deterministic mouse-motion profiles.

• MCP servers
• Browser automation

A benchmark and paper repository for evaluating Graph Retrieval-Augmented Generation (GraphRAG) models against traditional RAG, testing hierarchical knowledge retrieval and deep contextual reasoning.

• RAG
• AI evals

A DeepSeek Harness plugin that adds a full AI image workflow to the chat interface: text-to-image, continuous editing, a drawing canvas, batch generation, model comparison and a local ComfyUI connection.

• Agent skills
• Generative media

A Python library that provides an actor-model framework for connecting AI agents with traditional software systems, so developers can build custom agent-integrated applications on a minimal foundation.

• Agent frameworks

Ardublockly is a visual, drag-and-drop programming editor for Arduino boards, built on Google's Blockly, that generates Arduino C++ code and can upload it directly to a connected board.

• Editors

A CLI that launches an AI coding agent inside a dedicated sandboxed VM, local or in the cloud, bringing along existing skills, plugins and settings for Claude Code, Codex or OpenCode, with a browser, screen sharing and a warmed-up IDE.

• Coding agents
• Containers

source_gen is a framework and a convention for writing Dart source code generators, so multiple generators from different packages can produce code into the same project without clashing.

• Libraries

A JetBrains-branded desktop and mobile client for the `pi` coding agent that runs `pi` in-process and bridges it to a mobile-first UI, opening a Git repository as a project with per-worktree workspaces.

• Coding agents

A tool that lets a coding agent compose, generate and edit video from a plain-language description, expressing the result as an editable, re-renderable plan rather than a black-box render.

• Generative media

A source-available control plane that enforces policy and requires human approval before an AI agent's risky tool calls, shell commands, workflows or production changes, keeping an auditable record of what was approved.

• AI evals
• Security

An experimental platform for long-horizon workflow automation with teams of AI assistants, supporting OpenAI, Google, Anthropic and self-hosted open-weight models through a visual workflow editor and MCP tools.

• Agent frameworks
• Workflow automation

This repository publishes the paired coding exercises the Guardian newspaper's engineering recruitment process uses, so developers preparing for an interview can see the actual exercises and how they are run.

• Tutorials

The customer-run client for OpenAI's Secure MCP Tunnel, connecting a private or localhost MCP server to ChatGPT, Codex, the Responses API and AgentKit through an OpenAI-hosted tunnel endpoint, without exposing the server to the public internet.

• MCP servers
• Networking

A single local binary that gives AI agents persistent, cross-session memory as a graph-based store, with intent-aware recall, importance decay and automatic deduplication, supervised by the host LLM rather than an embedded model.

• Agent memory