One memory for OpenClaw, and every tool you will use
Give OpenClaw the Hypermemory CLI and skill so your agent reads and writes the same graph you use in Claude, n8n, and the desktop app.
Why This Works
Hypermemory is accessed via MCP and the CLI, the same memory graph either way. OpenClaw typically uses the CLI and skill with your API key and graph ID; other tools can use MCP on the same graph. Store once in messaging apps, recall in ChatGPT or Claude without re-explaining the project.
Quick Setup
- Download the Hypermemory CLI.
- Get an API key and graph ID from inside Hypermemory (app.hypermemory.io → API Keys).
- Give OpenClaw your API key and graph ID.
- Ask OpenClaw to download and use the Hypermemory skill. Now OpenClaw has access to the memory.
Available Tools
| Tool | Description |
|---|---|
| hm_find_related | Traverse connected knowledge from a starting node |
| hm_forget | Delete a node with optional cascade |
| hm_get_overview | Graph stats, top nodes, and type distribution |
| hm_ingest | Decompose dense text into graph entities in one pass |
| hm_list_files | Query uploaded files by type or name |
| hm_list_orphans | List disconnected nodes for cleanup |
| hm_recall | Hybrid search across your graph (BM25, vector, edges) |
| hm_store | Create a memory node; background enrichment runs automatically |
| hm_timeline | Temporal recall by period or date range |
| hm_timeline_write | Write an explicit diary-style timeline entry |
| hm_update | Update a node; re-embeds when the description changes |
| hm_upload_file | Upload a file into memory (summary + graph node) |
Use Case
Your OpenClaw assistant runs on messaging apps all week. Decisions and contacts land in Hypermemory. Friday you open Claude for a write-up, it already knows the client, the hold-firm pricing call, and who said what. You did not copy files between tools.