Overview
lucasastorian/llmwiki is a full-stack LLM Wiki implementation with 1.3k GitHub stars. It is the most feature-rich implementation, with a hosted option, Chrome extension, email integration, and course mode.
Key Facts
- Platform: Full-stack (Python 54%, TypeScript 44%)
- Stars: 1.3k
- Unique feature: Includes a hosted option, making it accessible without any local setup — the only major LLM Wiki tool with this capability
- Note: Not OKF compliant. Its full-stack architecture means it is the only implementation that works entirely in a browser with no local installation.
Features
Core Architecture
- Full-stack web app — Python backend + TypeScript frontend, works entirely in browser
- Hosted option — zero local setup required; unique among LLM Wiki implementations
- PostgreSQL + pgvector — vector storage for semantic search
- Background workers — Celery task queue for async processing
Ingest & Processing
- Chrome extension — web clipping with full-page or selected-content capture
- Email-to-save — forward emails to auto-ingest into the wiki
- Public ingest API — webhook endpoints for programmatic ingestion
- Batch processing — multiple sources at once
Search & Retrieval
- Semantic search — pgvector-powered vector similarity search
- Hybrid search — combines keyword and semantic matching
- Filtering — by tags, types, and custom metadata
MCP / Tools
- MCP server — Model Context Protocol integration for Claude.ai
- Zapier integration — workflow automation hooks
- Webhooks — event-driven integrations
Quality & Governance
- Course mode — progress tracking for educational content
- Public sharing — share wiki pages publicly
- Tagging system — organize knowledge with custom tags
Integration
- Claude.ai MCP — direct integration via Model Context Protocol
- Chrome extension — browser-based capture and save
- Email integration — forward-to-save workflow
- API access — programmatic access for automation
Data & Storage
- PostgreSQL + pgvector — relational + vector database
- Cloud-hosted — no local storage required
- Export — markdown export available
Comparison with Other Implementations
lucasastorian is the most feature-rich but also the most complex to self-host:
| Feature | lucasastorian | nashsu | atomicstrata |
|---|---|---|---|
| OKF Compliant | No | No | Yes |
| Interface | Full-stack web | Desktop GUI | CLI + SDK |
| Hosted Option | Yes | No | No |
| Web Clipper | Yes | No | No |
| Course Mode | Yes | No | No |
This makes it the best choice for users who want a hosted, full-featured knowledge base with web clipping and workflow automation, while nashsu-llm-wiki is simpler for a standalone desktop experience and atomicstrata-llm-wiki-compiler is better for standards compliance.
Relationships
- llm-wiki — implemented-by: Feature-rich implementation of the LLM Wiki pattern
- nashsu-llm-wiki — sibling-of: Alternative popular implementation (desktop app)
- atomicstrata-llm-wiki-compiler — sibling-of: OKF-compliant alternative
- claude-code — compatible-with: Can be used alongside for local wiki operations (llm-wiki-agent, claude-obsidian)
- hjarni — sibling-of: another hosted MCP wiki option
- capacities — sibling-of: object-based second-brain alternative
- swarmvault — sibling-of: local-first wiki + graph service
🔗 Conexão inferida — sugestão do segundo cérebro
Liga para: llm-wiki, andrej-karpathy Evidência observável: Lucas Astorian built a personal LLM Wiki following Karpathy’s pattern; his implementation is cited as a case study in the AI community.