KiwiFS: A Markdown Filesystem for AI Agents and Teams

KiwiFS is an open-source Markdown filesystem designed for AI agents and human teams, offering features like search, versioning, and structured data capabilities. It aims to bridge the gap between human-readable Markdown files and the structured data needs of AI agents.

RDR88Confidence 85%AI agentsfilesystemmarkdownknowledge baseversion controlopen sourceGo

Why it matters

KiwiFS addresses a common challenge in AI development: enabling agents to interact with and manage knowledge bases in a human-readable and machine-interpretable format. By making Markdown files searchable, versioned, and queryable, it could streamline the development of AI agents that require robust knowledge management and collaboration features.

KiwiFS is a Markdown filesystem built in Go, providing a searchable, structured, and versioned environment for AI agents and teams. It allows for native integration with various AI clients, including Claude and Cursor, through its 62 MCP tools. The project features a built-in web UI with wiki links, backlinks, and a knowledge graph. For data retrieval, it supports full-text search via SQLite FTS5 and pluggable vector search with options like OpenAI, Ollama, Cohere, Qdrant, and Pinecone. All data is versioned using Git, with every write resulting in an atomic commit, enabling features like blame, diff, and point-in-time restore. Users can perform SQL-like queries over frontmatter using DQL. KiwiFS offers 6 access protocols (REST, MCP, NFS, S3, WebDAV, FUSE) and 19 data importers for various sources including Postgres, MongoDB, Notion, and Obsidian. It also includes content health features like stale page detection and broken link identification, supports multi-space environments, webhooks, and schema validation with JSON Schema. The project is self-hostable and can be embedded as a Go library.

Article ID - cmpxd46gr0Featured on AI Radar: KiwiFS: A Markdown Filesystem for AI Agents and Teams