Explore the directory

Persistent memory for Claude Code that captures decisions, conventions and preferences during a session and recalls them later by meaning, with sub-100-millisecond semantic search, instead of relying on a single static CLAUDE.md file.

• Agent memory

An open-source, local-first alternative to NotebookLM that lets a person chat with their own PDFs, images, audio, video, URLs and YouTube links using whichever LLM provider they choose, including fully local models.

• RAG

A personal knowledge base where an LLM restructures raw notes and sources into a linked, structured wiki of Markdown articles rather than storing every word for later semantic search.

• Agent memory

A personal collection of LLM system prompts, agentic personas and prompt-engineering experiments, including a set of prompts for guiding image-to-video generation from an existing image using vision-capable models.

• Agent skills

A self-hostable runtime for orchestrating multiple AI agents that natively supports the MCP, A2A and AG-UI protocols, letting agents call outside tools, interoperate with agents built on other frameworks, and stream their state to a frontend in real time.

• Agent frameworks

A free, French-language series of 22 notebook-based chapters that build a language model entirely by hand, from basic vector arithmetic through attention, the transformer architecture, and a reasoning agent.

• Tutorials

A local-first command-line tool that runs a task through a fixed pipeline of specialized AI agents, architect, coder, reviewer, security, tester, fixer and judge, repeating the loop until quality thresholds pass before shipping.

• Coding agents

A vector-native data lakehouse file format, written in Rust and built on the Apache Iceberg specification, that stores tabular data, embeddings and an HNSW vector index together in one self-contained Parquet-extended file.

• Databases
• Data engineering

An open specification for a structured file that a software project includes to declare which parts of its code were generated or assisted by AI, and to what degree, so that a reviewer knows exactly which parts to double-check.

• Code quality

A self-hosted platform for running an organization of AI agents that have defined roles, hand work to each other, and answer to policies over which tools they may call, what they may spend, and what a person must approve first.

• Agent frameworks

Reproducible OCI container images, built with Nix, that sandbox AI coding agents such as Claude Code, Codex CLI and OpenCode so that a compromised or overreaching agent has limited access to the host filesystem and credentials.

• Containers

An IDE-integrated coding assistant, built on IBM's Granite model, that offers chat-based code suggestions, single- and multi-line code completion, and commands to explain or document existing code across languages including Go, C, C++, Java, JavaScript, Python and TypeScript.

• Coding agents

A collection of plugins and a shared knowledge runtime that give coding agents such as Codex, Kimi Code and Claude Code domain context, stable remote tools and a feedback loop of accumulated experience for vLLM and vLLM-Ascend development tasks.

• Agent skills

A Solargraph plugin that adds RSpec-aware code completion, jump-to-definition and type inference for Ruby projects, covering RSpec's describe/it blocks, memoized let helpers, custom matchers and several popular RSpec extensions.

• Editors

An agentic command-line coding harness built entirely on Elixir and Erlang/OTP, originally aimed at DeepSeek models, using process isolation so a failing tool call cannot crash the whole interactive session.

• Coding agents

A detector and repair tool for Chinese text corruption, or mojibake, that AI coding assistants can silently introduce when editing files, usable as a GitHub Action, a Claude Code hook, or a standalone command-line scan.

• CLI tools
• Code quality

The source code for a paper on a token-efficient, agentic retrieval-augmented generation framework that compresses both retrieved content and reasoning steps to answer complex queries with fewer tokens and faster reasoning.

• RAG
• Papers

The official code for an ECCV 2026 paper that generates realistic human motion sequences from text descriptions by combining retrieval of similar existing motions with a masked transformer generation model.

• Generative media
• Papers

A deterministic, formally specified orchestration tool for multi-agent systems, letting agents such as Claude and Codex be arranged into managed hierarchies and controlled from a web or terminal mission-control interface built on tmux.

• Agent frameworks

A training platform that repeats a four-phase cycle to build adversarial robustness into a text-classification model without eroding the defenses learned in earlier cycles, and generates EU AI Act compliance reports.

• Models
• AI evals

A Korean-language collection of guides, tutorials and hands-on labs that teach running local LLMs, building retrieval-augmented AI over personal documents, using AI coding agents responsibly, and designing agent skills, MCP tools and context engineering.

• Tutorials

A small, unofficial MCP server built on Jina AI's Reader and Search APIs that gives a coding agent exactly two tools: finding a short list of relevant web pages and reading the ones that matter.

• MCP servers

A method that writes new facts directly into the n-gram lookup table some recent language models carry alongside their transformer, as a small removable overlay file, without changing any of the model's trained weights.

• Models

A local-first runtime that gives every AI coding tool on a machine, such as Cursor, Claude Code, Copilot and Aider, one shared encrypted memory, safety layer and live session bus, kept per project and never committed to git.

• Agent memory