Connect a self-hosted MemOS service to DeepSeek Harness through the stock MCP client and expose persistent-memory tools.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-10-31
License
MIT
01What can it help you accomplish?
Use persistent memory from a DSH agent
Add, search, inspect, update and delete MemOS memories
DeepSeek Harness users running their own MemOS server
02How to install into DeepSeek Harness
Prerequisites
- Installed dsh CLI and Python ≥ 3.10.
- A MemOS checkout with Neo4j, Qdrant and the MemOS API running; reachable LLM and embedding gateway.
Installation steps
- 01
From the plugin checkout, run ./setup.sh --memos /path/to/MemOS on POSIX, or .\setup.ps1 --memos C:\path\to\MemOS on Windows.
- 02
$ dsh plugin --profile web add ./dsh-memos-bridge
- 03
Point MEMOS_PYTHON at the MemOS virtualenv Python and MEMOS_HOME at the MemOS checkout as needed.
- 04
Restart the Harness with dsh --profile web.
Verify the integration
- Run dsh --profile web --dump-config and look for id: memos-mcp.
- After restart, tools appear as mcp__memos__*.
03DSH integration and capability boundaries
Native DSH configuration bundle mounting MemOS over MCP stdio.
Add memories
Create memories from text, documents or conversation messages.
Semantic memory search
Search across the user’s memory cubes.
Inspect and maintain memories
Inspect, update or delete individual memories.
04Who is it for? When not to use it?
Good for
- DeepSeek Harness users running their own MemOS server
Not for
- The MCP command runs as trusted executable code outside the agent sandbox; connect only to MemOS servers you run yourself.
- Source patches were tested against MemoryOS 2.0.30; other source states may require manual adaptation.
- Neo4j is a hard startup dependency; an unavailable LLM or embedding gateway also affects search and chat.
05Compatibility, maintenance and safety notes
- The MCP command runs as trusted executable code outside the agent sandbox; connect only to MemOS servers you run yourself.
- Source patches were tested against MemoryOS 2.0.30; other source states may require manual adaptation.
- Neo4j is a hard startup dependency; an unavailable LLM or embedding gateway also affects search and chat.
README claims; not independently audited.
06Frequently asked questions
How does this bundle connect to DSH?
It configures the built-in @deepseek-ai/dsh-mcp-client to launch MemOS over stdio without changing Harness code.
How can I verify the DSH setup?
Check dsh --profile web --dump-config for id: memos-mcp, then restart; tools should appear as mcp__memos__*.
Does the DSH bundle include a complete MemOS server?
No. Prepare a MemOS checkout, its Neo4j/Qdrant/API infrastructure, Python environment, and reachable LLM and embedding gateway.
Does the MemOS command run inside the DSH agent sandbox?
No. The README calls it trusted executable code outside the agent sandbox and says to connect only to servers you run yourself.
07Related DSH workflows
rea
by morluto
Reverse engineer anything with agents, from app behavior down to native binaries.
ruflo
by ruvnet
🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
reactive-resume
by reactive-resume
A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!
everos
by evermind-ai
One portable memory layer for every AI agent: local-first, Markdown-native, user-owned, and self-evolving across apps, tools, and workflows.
08Data and sources
A [DeepSeek Harness](https://github.com/deepseek-harness/deepseek-harness) **bundle** that bridges the [MemOS](https://g…
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-10-31. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.