BotmemPersonal Memory for AI Agents

Your memories. Your agents. Your control.

🧠

Personal Memory

Ingest emails, messages, photos, and locations into a unified memory store. Everything you have said, received, or experienced — searchable and structured.

🔍

Semantic Search

Vector embeddings power natural-language queries across all your data. Ask "what did John say about the project deadline?" and get ranked, scored results.

🔌

Multi-Source

Gmail, Slack, WhatsApp, Telegram, iMessage, Immich photos, OwnTracks locations — and a Connector SDK to build your own.

🤖

Agent-Ready

Expose your memory to AI agents via REST API and CLI. Claude, GPT, and any agent can query your personal knowledge base.

🔒

Encrypted & Private

AES-256-GCM encryption at rest with a personal recovery key. Self-host on your hardware, or use the managed Pro tier (coming soon) — either way, your data stays yours.

🧩

Extensible

Build custom connectors with the Connector SDK. Add new data sources in under 200 lines of TypeScript.

Quick Start

Get Botmem running in under five minutes:

bash

# Clone the repository
git clone https://github.com/botmem/botmem.git
cd botmem

# Configure environment
cp .env.example .env    # Edit .env — set OLLAMA_BASE_URL to your Ollama host

# Start everything (Botmem + PostgreSQL + pgvector + Redis)
docker compose up -d

The API and web UI serve on http://localhost:12412. Sign up, connect your first data source, and start searching.

For development, see the quickstart guide.

Want someone else to handle the infrastructure? Botmem Pro ($14.99/mo, coming soon) will run the same open-source code with managed hosting and preconfigured AI.