# SynapseHRAG

> Hierarchical RAG with Hebbian Learning — Shared AI Memory. SynapseHRAG is a neural knowledge management MCP server that gives AI agents persistent, shared memory organized in 3 cognitive levels (strategic, tactical, operational) with Hebbian learning. 36 MCP tools, constitutional AI governance, multi-agent collaboration (Claude, ChatGPT, Codex, Gemini). Benchmarks: +118% context preservation, +36.5% precision, +40.4% recall vs standard RAG. MCP endpoint: https://hrag.synapsecorp.eu/mcp (Streamable HTTP). Company: SynapseCorp. Contact: ci@synapsecorp.eu.

## Documentation

- [Landing Page](https://hrag.synapsecorp.eu/)
- [API Reference](https://hrag.synapsecorp.eu/docs)
- [Privacy Policy](https://hrag.synapsecorp.eu/privacy)
- [User Dashboard](https://hrag.synapsecorp.eu/dashboard)
- [Full LLM Technical Reference](https://hrag.synapsecorp.eu/llms-full.txt)

## MCP Integration

- [MCP Endpoint](https://hrag.synapsecorp.eu/mcp)
- [OAuth2 Discovery](https://hrag.synapsecorp.eu/.well-known/oauth-protected-resource)
- [Authorization Server Metadata](https://hrag.synapsecorp.eu/.well-known/oauth-authorization-server)

## Optional

- [Sitemap](https://hrag.synapsecorp.eu/sitemap.xml)
- [Logo 128x128](https://hrag.synapsecorp.eu/logo-128.png)
- [Logo 512x512](https://hrag.synapsecorp.eu/logo-512.png)