mem9
mem9-ai/mem9
Persistent memory layer for AI agents with a Hermes Agent plugin, hosted API or self-hosted server
mem9 is a persistent memory service for AI agents that shares memory across sessions, machines and agent runtimes. Hermes Agent connects to it through a separate memory provider plugin, mem9-hermes-plugin.
What mem9 does
mem9 stores, searches, gets, updates and deletes memories through a server API, so agent plugins stay thin while storage, search, ingest and policy live on the server. Recall is hybrid, combining semantic and keyword search, and a visual dashboard lets you inspect what is stored. You can use the hosted API at api.mem9.ai or run the self-hosted mnemo-server, which takes a MNEMO_DSN database connection string.
The same memory can be shared by OpenClaw, Hermes Agent, Claude Code, OpenCode, Codex, DeepSeek Harness, Dify apps and custom HTTP clients. For Hermes Agent the integration is a memory provider plugin with a setup and activation flow, documented in the separate mem9-hermes-plugin repository. The repository description frames mem9 as memory for OpenClaw, and OpenClaw installs are the first path in the README.
Key features
- Persistent memory across sessions and machines
- Shared memory for multi-agent workflows across runtimes
- Hybrid semantic and keyword recall
- Visual dashboard for inspecting stored memories
- Hosted API with instant space provisioning, or a self-hosted mnemo-server
- Memory provider plugin for Hermes Agent with a setup and activation flow
When to use it
- Giving Hermes Agent memory that survives restarts and a switch to another laptop
- Letting Hermes Agent and Claude Code recall the same project facts
- Self-hosting memory storage on your own database
Who it is for: People running one or more coding agents who want shared, persistent memory across tools and machines.
How it fits with Hermes Agent
Supports Hermes Agent among several runtimes through the separate mem9-hermes-plugin memory provider. The main repository is a general memory server whose description focuses on OpenClaw.
Requirements: MEM9_API_URL and MEM9_API_KEY environment variables; self-hosting also needs a MNEMO_DSN database connection for mnemo-server
FAQ
What is mem9?
mem9 is a persistent memory service for AI agents. It keeps memory across sessions and machines and lets several agent runtimes recall the same facts through one server API.
Does mem9 work with Hermes Agent?
Yes, Hermes Agent is a listed runtime. The integration is a memory provider plugin with a setup and activation flow, described in the mem9-hermes-plugin README.
What do I need to run mem9?
You need a mem9 endpoint, either the hosted API at https://api.mem9.ai or a self-hosted mnemo-server, plus an API key set as MEM9_API_KEY. Self-hosting also needs a MNEMO_DSN database connection string.
Similar memory for Hermes Agent
All memoryPortable .agent/ folder of memory, skills and protocols that moves between coding-agent harnesses
MemTensor MemmyLocal memory hub and personal agent that shares one memory across Claude Code, Codex, OpenClaw and Hermes
activeloopai HivemindShared agent memory that turns session traces into reusable skills, with a Hermes install option
ClaudioDrews Memory OSSeven-layer local memory system for Hermes Agent with Qdrant, structured facts and a curated wiki
AVIDS2 MemorixLocal-first shared project memory for coding agents over MCP, including Hermes Agent
lout33 Symbiotic AIPlain Markdown files that give an AI agent persistent context about you, with an onboarding skill
Related guides: SOUL.md for Hermes Agent: what it is and how to write one · Run multiple Hermes agents with profiles