Supermemory for OpenClaw
https://github.com/supermemoryai/openclaw-supermemory- Category
- AI Agents
- Rank
- No. 996Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- supermemoryai
- GitHub
- 799 stars
- Date
About
Long-term memory and recall for your OpenClaw agent. Connect Supermemory so your bot remembers context across sessions.
What it does
The plugin wraps each agent turn with two automatic steps. It fetches profile facts and semantically related conversations before a response, then submits the completed exchange for extraction and storage. Manual commands and agent tools also support saving, searching, inspecting, and deleting memories.
Why it's ranked here
This is a focused, practical integration with useful control points. Automatic recall and capture reduce routine work, while configurable limits, profile frequency, capture filtering, namespaces, and deletion tools keep the system manageable. Its value depends on accepting Supermemory as the memory backend and granting conversation access.
What's good
Setup is short, yet advanced configuration exposes meaningful choices. Users can disable either automatic hook, separate work and personal memories, select a custom endpoint, and inspect or wipe stored data. Recall output deduplicates profile facts and search results, includes relevance information, and avoids running for noninteractive triggers.
Tradeoffs
The hosted service requires a Pro plan or higher, and automatic capture sends conversations to Supermemory after every eligible turn. Recall injects external context into prompts, so privacy and prompt trust deserve attention. Query-based forgetting deletes the closest search match, which is less exact than deleting a known memory identifier.
How to use it well
Use it for an OpenClaw assistant that repeatedly needs personal preferences, decisions, projects, or prior conversation context. Start with filtered capture, review the profile and search results, then add containers when domains genuinely need separation. It does not provide a local memory-file workflow or a general knowledge store outside OpenClaw.
Technical notes+
package.json defines an MIT-licensed ESM npm package, @supermemory/openclaw-supermemory, with TypeScript sources, an esbuild Node ES2022 bundle, a Supermemory SDK dependency, and an OpenClaw peer requirement of >=2026.5.7. index.ts registers the memory runtime, CLI, slash commands, agent tools, recall and capture hooks, plus compatibility aliases for tool names. config.ts rejects unknown configuration keys, resolves environment variables, validates endpoint protocols, and defaults to the cloud API. client.ts sanitizes content, validates API-key format, tags writes by source, supports profile search and bulk wiping, and permits a custom base URL. hooks/recall.ts limits recall to interactive triggers, strips recognized inbound metadata, deduplicates context, and catches remote failures. One maintenance concern is visible: package.json reports version 2.1.15 while index.ts uses 2.1.14 for update checks.
Observed
- License
- MIT
- Primary language
- TypeScript
- Packaging
- ESM npm package published as @supermemory/openclaw-supermemory
- Install surface
- Installed through the OpenClaw plugin installer
- Interfaces
- OpenClaw CLI commands, slash commands, agent tools, and lifecycle hooks
- Platform support
- Requires OpenClaw and plugin API version 2026.5.7 or newer
- Runtime target
- Bundled for Node.js with an ES2022 target
- Backend options
- Defaults to Supermemory cloud and accepts a custom HTTP or HTTPS base URL
Read from README.md, package.json, index.ts, lib/validate.js, lib/validate.d.ts, lib/version-check.ts, client.ts, config.ts, logger.ts, memory.ts, runtime.ts, tools/store.ts, commands/cli.ts, hooks/recall.ts, tools/forget.ts.
What it can do
Store conversation context across sessions
Chat history and conversation data → Persistent memory storage
Retrieve relevant past interactions
Current conversation context or query → Related historical conversation snippets
Maintain user preferences and settings
User configuration data and behavioral patterns → Personalized agent responses
Connect memory system to OpenClaw agent
OpenClaw agent instance and memory configuration → Integrated bot with memory capabilities
Recall previous task outcomes and results
Task identifiers or similar current tasks → Historical task results and learned patterns
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.