remind

Overview

Remind functions as a sophisticated memory layer for AI agents, offering both pattern concepts for generalizations and fact clusters for verbatim details. Unlike traditional RAG systems, it extracts generalized concepts from experiences through a consolidation process. The MCP server provides centralized memory for IDE agents like Cursor and Claude Desktop, featuring web UI, REST API, and multiple database backends including SQLite and PostgreSQL with vector search capabilities.

Try asking AI

After installing, here are 3 things you can ask your AI assistant:

• remember

  Add an episode to memory with optional type, entity, and metadata

  Try:Remember that this project uses React with TypeScriptTry:Remember that we chose PostgreSQL for the database with type decisionTry:Remember this important conversation about product roadmap
• recall

  Retrieve memories using semantic search or entity filtering

  Try:What tech stack are we using for this project?Try:Show me all decisions we've made about database choicesTry:Recall any conversations about user authentication
• consolidate

  Run memory consolidation to extract generalized concepts

  Try:Consolidate my recent memoriesTry:Run consolidation in the backgroundTry:Forcefully consolidate all memories now
• ingest

  Auto-ingest raw text to extract memory-worthy episodes

  Try:Ingest this raw conversation transcriptTry:Process this meeting transcript for memory extractionTry:Ingest the Slack channel history
• inspect

  List or view details of concepts or episodes

  Try:Inspect all concepts about database decisionsTry:Show me the most recent episodesTry:Inspect details of concept with id 123
• search

  Search across concepts using keywords or tags

  Try:Search for anything related to authenticationTry:Find concepts tagged with 'performance'Try:Search for mentions of 'TypeScript'
• entities

  List entities or show details of a specific entity

  Try:Show all entities in this projectTry:Show details for the entity ReactTry:List entities mentioned in recent episodes
• topics

  Manage topics and view topic-specific concepts

  Try:List all topicsTry:Create a new topic called 'Performance optimizations'Try:Show top concepts for the product roadmap topic
• update-episode

  Update content, type, entities, or topic of an episode

  Try:Update episode 45 with more detailsTry:Change the type of episode 23 to decisionTry:Add entity 'React' to episode 12
• update-concept

  Update title, summary, confidence, tags, or relations of a concept

  Try:Update the title of concept 78Try:Add relation to concept 45Try:Change the confidence score for concept 12
• export

  Export memory to JSON format

  Try:Export all my memories to JSONTry:Export memories with format compactTry:Export memories with full episode details
• import

  Import memories from JSON format

  Try:Import memories from this JSON fileTry:Import memories from backup.jsonTry:Import exported memories from another project

Comparable tools

Installation

# Install the package
pip install remind-mcp

# Start the MCP server
remind-mcp --port 8765

{
  "mcpServers": {
    "remind": {
      "url": "http://127.0.0.1:8765/sse?db=my-project"
    }
  }
}

pip install "remind-mcp[postgres]"  # For PostgreSQL support
pip install "remind-mcp[rerank]"   # For cross-encoder reranking

Compare remind with