Personal Knowledge Manager

Part of the worlds-biggest-software-project initiative.

An AI-native second brain that captures, links, and retrieves your personal knowledge through semantic search and conversational RAG.

Personal Knowledge Manager is an open-source second-brain tool that turns scattered notes, meeting transcripts, web clips, and documents into a searchable personal knowledge graph. It is built for researchers, knowledge workers, product managers, students, writers, and content creators who want AI-grade retrieval without surrendering their data to a closed cloud service.

Why Personal Knowledge Manager?

Notion AI feels bolted-on rather than native to the information model, and AI add-ons are billed separately on top of an already escalating per-seat price.
Obsidian gives complete data ownership but requires plugin configuration before AI features work, and its mobile editing experience still lags the desktop.
Mem auto-organises with AI but offers no Android app, no manual linking tools, and no public developer API.
Tana and Capacities collapse notes and databases into one model but impose steep learning curves and, in Tana's case, cap the free plan at 5 supertags.
Roam Research pioneered bi-directional linking but has stagnated, has no native AI, and charges $15/month against free or cheaper rivals.

Key Features

Capture and Editing

Markdown note editor with YAML frontmatter and bi-directional [[WikiLink]] support
Daily note / journal as the primary capture surface
Web clipper browser extension for sending pages into the vault
Voice memo ingestion: transcribe and classify into the knowledge graph automatically

AI Retrieval and Synthesis

Semantic embedding of notes on save, with related notes surfaced during editing
Conversational retrieval (RAG chat) over the personal corpus with citations
Full-text and semantic search side by side
Proactive synthesis on demand across hundreds of scattered notes

Graph and Structure

Graph view of note connections
Bi-directional linking with automatic backlinks
Optional object / typed-note model for structured knowledge (similar to Capacities and Tana)
Spatial whiteboard canvas for visual thinkers

Interoperability and Agents

Import from Obsidian, Notion export, and plain Markdown files
MCP server endpoint exposing the vault to AI agents
Spaced-repetition review scheduling for notes and concepts
Real-time collaboration on shared vaults

AI-Native Advantage

Embedding-based linking removes the manual [[link]] tax: relevant prior notes surface automatically as you write. A grounded chat interface answers questions over your own corpus with citations, turning the vault into a personal research assistant. Capture-anywhere agents — voice memos, email forwarders, web clippers, meeting transcripts — extract and classify knowledge into the graph without context-switching, and forgetting-curve scheduling keeps important notes resurfacing on a decay-aware timetable.

Tech Stack & Deployment

The project targets a local-first, self-hostable deployment with optional cloud sync, in line with the privacy-first direction of the wider PKM market and the trend toward private RAG over locally stored documents. Notes are stored as plain CommonMark Markdown with YAML frontmatter for portability. Vector embeddings (OpenAI or Sentence Transformers class models) power semantic search; OAuth 2.0 connects external sources such as email, calendar, and Drive; WebDAV is the baseline file-sync standard. An MCP server endpoint provides AI-agent access to the vault.

Market Context

The personal knowledge base AI market was valued at approximately $1.65 billion in 2025 and is projected to reach $2.16 billion in 2026 at a 30.3% CAGR, expanding to $6.15 billion by 2030 (Research and Markets, 2026). Incumbent individual pricing clusters at $8–$15/month (Notion AI add-on $10, Mem $14.99, Roam $15, Heptabase $11.99), with team plans at $15–$25/user/month. Primary buyers are researchers and academics, knowledge workers, product managers, students, and writers managing source material over years.

Project Status

This project is in the research and specification phase.
Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.