remembrallmcp

Overview

RemembrallMCP is a comprehensive MCP server that addresses the stateless nature of AI coding agents by providing persistent memory and code dependency analysis. It uses Rust for performance, PostgreSQL with pgvector for vector storage, and implements the MCP protocol to expose tools for memory recall/management and code intelligence. The server significantly reduces token usage by pre-building dependency graphs rather than having agents explore codebases from scratch each session.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• remembrall_recall

  Search memories - hybrid semantic + full-text with RRF fusion

  Try:Recall what we know about database decisionsTry:Find information about authentication patternsTry:What did we decide about caching strategy?
• remembrall_store

  Store decisions, patterns, knowledge with vector embeddings

  Try:Store: We chose Postgres over MongoDB because our query patterns are relationalTry:Remember this decision about the API designTry:Document this pattern for future reference
• remembrall_update

  Update an existing memory (content, summary, tags, or importance)

  Try:Update the memory about database decisions with more contextTry:Change the importance score for that architectural decisionTry:Add tags to this memory about authentication
• remembrall_delete

  Remove a memory by UUID

  Try:Delete the memory with UUID 123e4567-e89b-12d3-a456-426614174000Try:Remove that outdated decision about the old APITry:Get rid of the incorrect information about the database schema
• remembrall_ingest_github

  Bulk-import merged PR descriptions from a GitHub repo

  Try:Ingest PRs from myorg/myrepo repositoryTry:Import the last 100 merged PRs from our projectTry:Fetch GitHub PR history for this codebase
• remembrall_ingest_docs

  Scan a directory for markdown files and ingest them as memories

  Try:Ingest all markdown files from the docs directoryTry:Process the project README and architecture docsTry:Scan the ADRs folder for documentation
• remembrall_index

  Parse a project directory into a dependency graph (8 languages)

  Try:Index the current project directoryTry:Build the dependency graph for this codebaseTry:Parse the source code to understand relationships
• remembrall_impact

  Blast radius analysis - what breaks if I change this?

  Try:Show me the impact of changing UserServiceTry:What files are affected if I modify this function?Try:Find all callers of the AuthMiddleware
• remembrall_lookup_symbol

  Find where a function or class is defined across the project

  Try:Where is the UserService class defined?Try:Find the implementation of authenticate functionTry:Look up the DatabaseConnector symbol

Comparable tools

Installation

git clone https://github.com/cdnsteve/remembrallmcp.git
cd remembrallmcp
# Start Postgres + initialize schema + download embedding model
docker compose up -d
# Run the MCP server
docker compose run --rm remembrall

# macOS (Apple Silicon)
curl -fsSL https://github.com/cdnsteve/remembrallmcp/releases/latest/download/remembrall-aarch64-apple-darwin.tar.gz | tar xz
sudo mv remembrall /usr/local/bin/

# Initialize
remembrall init

{
  "mcpServers": {
    "remembrall": {
      "command": "remembrall"
    }
  }
}

FAQ

What languages are supported by the code dependency graph?

Currently supports 8 languages: Python, Java, JavaScript, Rust, Go, Ruby, TypeScript, and Kotlin with quality scores ranging from A to B.

How does this improve AI agent performance?

By reducing token usage by 95-98% for code exploration tasks. Instead of agents reading multiple files per query, RemembrallMCP provides instant answers through pre-built dependency graphs and memory recall.

Compare remembrallmcp with