Open-source memory server for LLM chat applications, built in Rust with Redis storage and automatic conversation summarization.
A simple memory server for AI chatbots that stores conversation history and auto-summarizes old messages using Redis and OpenAI.
MotorHead is an open-source memory server from Metal that does one thing: store and manage conversation history for LLM chat applications. It runs as a Rust binary (or Docker container), backed by Redis, and exposes a REST API with three core operations: post messages, get context, delete sessions.
The main trick is sliding window management with incremental summarization. You set a window size (say, 20 messages). When the conversation exceeds that, MotorHead calls OpenAI to summarize older messages into a compressed "long-term memory" block. New messages update the summary incrementally rather than regenerating from scratch, which keeps latency and API costs low during long conversations.
Deploy it with Docker Compose and you're running in under five minutes. The Redis backend handles thousands of concurrent sessions with sub-millisecond reads. Sessions get isolated storage with configurable TTL for automatic cleanup. For teams already running Redis, MotorHead adds minimal operational overhead.
The LangChain integration (both Python and JS) works out of the box, though the LangChain docs note this integration is deprecated as of v1.0 (October 2025). You can still use the REST API directly from any language.
Here's where it falls short: MotorHead only does linear conversation recall. No semantic search across past conversations, no entity extraction, no knowledge graphs, no cross-session memory. If a user mentions their dog's name in session 1, session 2 won't know about it. Tools like Mem0 and Zep handle those cases. MotorHead doesn't try to.
Maintenance has slowed considerably. The GitHub repo (907 stars, Apache-2.0 license) shows sparse commits since 2023, and Metal has shifted focus to other products. The server works, but expect no significant feature development or rapid bug fixes. For new projects in 2026, Mem0 or Zep are safer long-term bets. MotorHead remains useful if you need something minimal that you deploy once and leave running.
Was this helpful?
MotorHead does one thing well: it stores chat messages and auto-summarizes old ones so your LLM has context. The Rust+Redis combo is fast, deployment is trivial, and the REST API works from any language. But it stops there. No semantic search, no cross-session memory, no entity awareness, and maintenance has stalled. Good for simple chatbot memory needs; look at Mem0 or Zep if you need anything more.
Maintains a configurable window of recent messages. When exceeded, older messages are compressed into a running summary rather than dropped. Default window is 12 messages but configurable via environment variable.
Use Case:
A customer support chatbot keeps the last 20 messages in full while preserving a summary of the entire conversation for context, so the agent doesn't repeat questions already answered.
Updates the conversation summary as new messages arrive instead of regenerating from scratch. Each summarization call only processes the new messages against the existing summary, reducing OpenAI API costs by 60-80% compared to full re-summarization.
Use Case:
A 200-message therapy bot session where the summary updates in real-time without reprocessing the entire history each turn.
All session data stored in Redis with configurable TTL for automatic cleanup. Leverages Redis's in-memory speed for sub-millisecond read/write operations. Works with any Redis instance, including managed services like AWS ElastiCache.
Use Case:
A SaaS platform serving 5,000 concurrent chat sessions needs sub-millisecond memory retrieval without managing a separate database.
Three endpoints: POST messages to a session, GET context (recent messages plus summary), DELETE sessions. No framework dependencies. Any language that makes HTTP requests works.
Use Case:
A Go or Rust backend integrates chat memory without pulling in Python, LangChain, or any AI framework.
Available as a Docker image with included Docker Compose configuration for the full MotorHead + Redis stack. Single command brings up both services.
Use Case:
A developer prototyping a chatbot deploys persistent memory in under 5 minutes with docker-compose up.
Free
Ready to get started with MotorHead?
View Pricing Options →MotorHead works with these platforms and services:
We believe in transparent reviews. Here's what MotorHead doesn't handle well:
Weekly insights on the latest AI tools, features, and trends delivered to your inbox.
AI Memory & Search
Mem0: Universal memory layer for AI agents and LLM applications. Self-improving memory system that personalizes AI interactions and reduces costs.
AI Memory & Search
Context engineering platform that builds temporal knowledge graphs from conversations and business data, delivering personalized context to AI agents with <200ms retrieval latency.
AI Memory & Search
Open-source framework that builds knowledge graphs from your data so AI systems can analyze and reason over connected information rather than isolated text chunks.
AI Memory & Search
PostgreSQL-native vector search via pgvector integrated into Supabase's managed backend — store embeddings alongside your relational data with auth, real-time subscriptions, and row-level security.
No reviews yet. Be the first to share your experience!
Get started with MotorHead and see if it's the right fit for your needs.
Get Started →Take our 60-second quiz to get personalized tool recommendations
Find Your Perfect AI Stack →Explore 20 ready-to-deploy AI agent templates for sales, support, dev, research, and operations.
Browse Agent Templates →