Explore the directory

A manual of practical engineering references covering AWS, Kubernetes and AI platforms, including EKS best practices, an agentic AI platform architecture guide, and infrastructure benchmark comparisons.

• Books

A local tool that indexes documents from Jira, Confluence or local files into a vector database and searches them with hybrid vector and keyword search, keeping all data on the user's own machine.

• MCP servers
• RAG

A practitioner's guide to Claude Code covering setup, CLAUDE.md, permissions, skills, hooks, sub-agents and MCP servers, with interactive terminal demos on every page.

• Tutorials
• Books

A bot that runs Claude Code sessions on a developer's own machine and streams them live into Slack or Mattermost threads, so a team can watch and steer a session together.

• Coding agents

A full-screen terminal chat interface for OpenAI-compatible AI APIs, with character cards, personas, presets and early support for the Model Context Protocol.

• Inference

A Claude Code skill that replaces a long sequence of manual instructions with one command, scanning a project, generating code with security checks on every file, and committing the result.

• Agent skills

A local MCP server, written in pure Dart, that lets an AI client audit, clean up and generate code from a Flutter project's assets and localization files.

• MCP servers
• Mobile

An unofficial X (Twitter) client for autonomous agents, written in Go as a single binary and library, that signs in with a personal account's cookies and prints one JSON object per command.

• CLI tools
• APIs

An F# authoring layer over the Microsoft Agent Framework that adds a higher-level, typed and declarative workflow language for building durable agent pipelines on .NET.

• Agent frameworks

An independent, from-scratch implementation of the Kimi K3 language model architecture described in its paper, reproducing its published parameter counts and running the same code path from a 20-million-parameter laptop model up to its 2.8-trillion-parameter configuration.

• Models

The source for the documentation site covering Nefarius' projects on USB input-device emulation, game peripherals and reverse-engineering tooling, built with MkDocs Material.

• Documentation tools

A self-hosted operational database for auditing AI agent behavior, recording each step's causal chain back to the user message so a team can answer why an agent called a given tool.

• Databases
• Monitoring

An open-source desktop AI coding agent that connects directly to any model provider with the user's own API key, writing code, running commands and browsing the web to verify its own work.

• Coding agents

A provider-agnostic Node.js SDK for retrieval-augmented generation that has grown a second half: a context and memory layer that decides what a model sees on a given turn under a fixed token budget.

• Agent memory
• RAG

A native macOS app and inference engine that runs large mixture-of-experts language models on Apple Silicon by streaming only the needed experts from SSD instead of holding a whole model in memory.

• Inference

A minimal, hand-written LLM inference engine for the MiniCPM5-1B model, built in Rust to study how transformer inference works, with a terminal UI that visualizes generation as it happens.

• Inference
• Tutorials

A hosting platform for AI agents meant to persist over time, giving each one a Docker sandbox, a self-managed home directory, decaying journal memory, scheduled unprompted heartbeats, and access to connected services such as Telegram and GitHub.

• Agent frameworks

A research framework that evolves a single meta-skill file describing how an LLM should design and run a multi-agent system, refining it over rounds using contrastive reflection on past attempts.

• Agent skills
• Papers

A per-project semantic search tool over a developer's own Claude Code session history, so a new session can recall what was decided in past ones instead of needing to be re-explained.

• Agent memory

A reinforcement-learning framework for repository-level code completion that trains a retriever to select reference code snippets that actually help generation, rather than merely similar ones.

• Models
• Papers

A webhook-driven service that runs automated AI code review on GitLab merge requests and GitHub pull requests through Claude Code, tracking progress on a real-time dashboard.

• Code quality

A set of retrieval-augmented-generation demos that run entirely offline on CPU with quantized open models such as DeepSeek, Llama, Qwen, Mistral and Gemma.

• RAG
• Tutorials

A collection of 50 standalone browser-based tools for analyzing, refining, scoring and building prompts for large language models, each a single HTML file with no build step.

• Agent skills
• AI evals

A research toolbox for training and evaluating multimodal large language models under continual instruction tuning, where a model learns new tasks in sequence without forgetting earlier ones.

• Models
• AI evals