engram

OneNomad-LLC/engram
★ 0 stars TypeScript 🤖 AI/LLM Updated today
Smart memory MCP server. Hybrid ANN search, spreading activation, temporal inference, knowledge graph, procedural rules.
View on GitHub →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "engram": {
      "command": "npx",
      "args": [
        "-y",
        "OneNomad-LLC/engram"
      ]
    }
  }
}

README Excerpt

A memory system for AI agents that actually works. LLMs can't remember anything between conversations by default, and the existing solutions are either too simple (just dump everything in a vector DB) or too expensive (send your entire history to an API every time). Engram sits in the middle. It runs locally, doesn't need an API key for basic operation, and scores 92.0% recall on the LoCoMo benchmark. That beats every open-source memory system I've tested against.

Tools (20)

ENGRAM_DATA_DIRENGRAM_DEVICEENGRAM_EMBEDDING_MODELENGRAM_MODELMEM0_API_KEYOPENROUTER_API_KEYmemory_context_pressurememory_diary_readmemory_diary_writememory_extractmemory_formatmemory_governmemory_handoff_readmemory_handoff_writememory_importmemory_ingestmemory_kg_addmemory_kg_invalidatememory_kg_querymemory_kg_timeline