Skip to content

Latest commit

 

History

History

docs

README.md

icon layout
hand-wave
width title description tableOfContents outline pagination metadata
default
visible
true
visible
visible
true
visible
true
visible
true
visible
true

Welcome to Almanac

Almanac is a lightning-fast data access platform designed specifically for AI agents. It combines graph-enhanced retrieval (LightRAG) with zero-config indexing to make any data source instantly accessible to your AI applications.

What Makes Almanac Different?

  • 🚀 Lightning Fast - Entity-based retrieval reduces tokens by 10x while improving accuracy
  • 🔌 Zero Config - Automatically generates indexing configurations for any MCP server
  • 🧠 Smart Retrieval - 5 query modes adapt to different use cases (naive, local, global, hybrid, mix)
  • 📊 Graph-Enhanced - Understands relationships between entities, not just keywords
  • ⚡ Production Ready - Parallel processing, multi-database architecture, built to scale

How It Works

Think of Almanac like a librarian who doesn't just know where books are, but understands how they relate to each other. When you ask a question:

  1. Syncing - Almanac fetches data from your sources (Slack, GitHub, Notion, etc.)
  2. Indexing - Creates both vector embeddings and knowledge graphs
  3. Query - Chooses the best retrieval strategy based on your needs
  4. Results - Returns relevant information with relationships and context

Quick Example

# Install and start (one command)
pnpm start

# Query your data
curl http://localhost:3000/api/query \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What did we discuss about the API refactor?",
    "mode": "mix"
  }'

Get Started in 5 Minutes

Ready to dive in? Follow our Quick Start Guide to:

  • Install Almanac with Docker
  • Connect your first data source
  • Run your first query
  • Understand the different query modes

Key Concepts

New to RAG or knowledge graphs? Start here:

Common Use Cases

See Almanac in action:

Why Developers Choose Almanac

"We tried building RAG from scratch. Almanac gave us better results in an afternoon than we achieved in 3 weeks."

— Dev team building AI code assistant

For Developers Building AI Agents:

  • No AI/ML expertise required
  • Works with any LLM (OpenAI, Anthropic, local models)
  • REST API - integrate with any stack
  • Full TypeScript codebase

For Data-Heavy Applications:

  • Handles millions of documents
  • 32 concurrent operations by default
  • Smart caching and batching
  • Vector + Graph + Document storage

Architecture at a Glance

External APIs → MCP Servers → Almanac
                                ↓
                    [Syncing & Transformation]
                                ↓
                    ┌─────────────────────┐
                    │     Databases       │
                    │  - MongoDB (docs)   │
                    │  - Qdrant (vectors) │
                    │  - Memgraph (graph) │
                    │  - Redis (cache)    │
                    └─────────────────────┘
                                ↓
                    [LightRAG Query Engine]
                                ↓
                            Results

Next Steps

Community & Support

LLM? Read llms.txt.

Built for developers, by developers. Open source and production-ready.