Explore the directory

An interactive C# REPL for the Unity Editor and Runtime, powered by Roslyn, with semantic code completion, cross-submission state, private member access and remote execution on an IL2CPP player build via HybridCLR.

• Debugging

A Visual Studio Code extension that adds AI pair programming through Editor Code Assistant (ECA), connecting to an external eca server for chat, code rewriting, inline chat and context management, with commands to sandbox the server process.

• Coding agents
• Editors

A terminal system monitor purpose-built for local LLM inference, showing live tokens per second, VRAM-spill and throttle warnings and other GPU metrics nvidia-smi does not expose, alongside a full htop or btop-class system monitor.

• Terminals
• Monitoring

Deployment configuration for self-hosting Dakera, an AI agent memory server that combines vector search, hybrid BM25 and HNSW retrieval, knowledge graphs and session management in a single Rust binary, covering Docker Compose, Kubernetes and Helm.

• Agent memory
• Containers

A thread-safe vector database for model inference embeddings, built in Rust on top of LMDB.

• Databases

A pure-PHP vector database implementing HNSW approximate nearest-neighbour search and BM25 full-text retrieval, which can be combined into a hybrid search pipeline.

• Databases

A self-hosted control plane above coding agents such as Claude Code, Cursor and OpenCode that routes tasks between them, tracks and limits spend, and runs multi-agent workflows.

• Coding agents
• Agent frameworks

A sparse fine-tuning method that trains a small fraction of a Mixture-of-Experts model's expert weights and router, making MoE fine-tuning possible on a 12GB consumer GPU.

• Models

A collection of integrations that connect TinyFish's natural-language web automation agent to AI assistants and workflow platforms, including bot-protected pages.

• Browser automation

A local-first C++ runtime for CPU-only LLM inference with Apple Silicon acceleration, exposed through a unified command-line interface.

• Inference

A set of personal notes and first principles for designing prompts, arguing prompts should behave like a programming language and that Python can serve as the prompt itself.

• Books

A benchmarking tool that shows exactly where a local LLM ran, CPU or GPU and on which layers, alongside prefill, decode, memory and power measurements.

• Inference
• Debugging

A PDF question-answering web app that lets users upload documents, ask questions grounded in their content, and generate summaries, through a local three-service stack.

• RAG

An open-source SystemVerilog NPU project building a bare-metal transformer accelerator for experimental Gemma-class LLM inference on an AMD/Xilinx FPGA board.

• Inference

A free, self-hostable, LLM-agnostic alternative to Anthropic's Claude Tag, giving a Slack channel one shared AI teammate instead of a personal per-user assistant.

• Agent frameworks

A privacy-focused, offline coding agent that runs entirely against a local Qwen model served by Ollama, with no network calls, accounts or API keys.

• Coding agents

A chat-driven LLMOps agent that turns a folder of documents into a fine-tuned, domain-specific language model, with human approval at each pipeline step.

• Models
• Data engineering

A long-running, self-evolving vulnerability-discovery harness that wraps any strong coding model in a structured recon-hunt-validate pipeline for security research.

• Security

A desktop-first AI coding agent that ships as an Electron GUI, a terminal TUI, a background daemon and an always-on desktop companion, sharing one agent engine.

• Coding agents

A self-hostable, omni-modal retrieval platform that searches documents, images, audio and video with hybrid retrieval, reranking and a bounded agent for follow-ups.

• RAG

A terminal coding agent that keeps a persistent memory of a repository's layout, recent work and corrected preferences, so sessions do not start from zero.

• Coding agents

A Dart plugin implementing a Model Context Protocol client, letting Flutter apps connect to MCP servers across Android, iOS, web, Linux, Windows and macOS.

• MCP servers

A CPU-first reference platform showing the systems around LLM inference: an OpenAI-compatible gateway, a deterministic eval CLI, an experiment registry and a RAG toolkit.

• Inference
• AI evals

A self-hosted web workspace for Google's Jules coding agent, adding live code-diff viewing, an integrated terminal and session analytics around the Jules API.

• Coding agents