OpenClaw plugin for Kybernesis Brain - infinite context and memory with hybrid search, auto-tiering, and sleep agent capabilities
npm install @kybernesis/openclaw-kybernesisOpenClaw plugin for Kybernesis Brain - infinite context and memory for AI agents with hybrid search, automatic tiering, and sleep agent capabilities.
- Hybrid Search: Combines vector similarity with metadata filtering for precise memory retrieval
- Automatic Tiering: Memories move between hot/warm/archive based on access patterns
- Sleep Agent: Background processing for auto-tagging, relationship detection, and summarization
- Multi-Source: Ingest from chat, file uploads, Google Drive, Notion, and more
- Cross-Platform: Works across all OpenClaw-compatible AI clients
``bash`
openclaw plugins install @kybernesis/openclaw-kybernesis
Add to your OpenClaw config:
`json`
{
"plugins": {
"@kybernesis/openclaw-kybernesis": {
"apiKey": "${KYBERNESIS_API_KEY}",
"autoRecall": true,
"autoCapture": true,
"maxRecallResults": 10
}
}
}
| Option | Type | Default | Description |
|--------|------|---------|-------------|
| apiKey | string | - | Your Kybernesis API key (or set KYBERNESIS_API_KEY env var) |mcpEndpoint
| | string | https://api.kybernesis.ai | Kybernesis MCP server URL |autoRecall
| | boolean | true | Inject relevant memories before each turn |autoCapture
| | boolean | true | Store important conversation content |maxRecallResults
| | number | 10 | Max memories to inject per turn (1-25) |includeTierInfo
| | boolean | false | Show memory tier in recalled context |captureMinLength
| | number | 50 | Min message length to trigger capture |captureTags
| | string | "openclaw" | Comma-separated tags for captured memories |debug
| | boolean | false | Enable verbose logging |
Save content to Kybernesis Brain.
`
Use this tool to remember important information.
Parameters:
- content (string): The content to store
- title (string, optional): Memory title
- tags (string[], optional): Tags for categorization
`
Search memories with hybrid vector + metadata search.
``
Parameters:
- query (string): Natural language search query
- limit (number, optional): Max results (default: 10)
Delete a memory from Kybernesis Brain.
``
Parameters:
- memoryId (string): ID of the memory to delete
Get memory statistics and context.
``
Returns total memories, distribution by tier and source, recent activity.
Manually save content to memory.
``
/remember The user prefers dark mode and uses TypeScript
/remember Meeting notes from standup tags:meeting,standup,daily
Syntax: /remember
Search your memories.
``
/recall user preferences
/recall TypeScript best practices limit:3
Syntax: /recall
Check connection and view statistics.
``
/kyber-status
When enabled, the plugin automatically searches for relevant memories before each AI turn and injects them as context. Memories are formatted in an XML-like wrapper that the AI can use naturally:
`xml
The following memories from Kybernesis Brain may be relevant...
1. User prefers dark mode
The user mentioned they always use dark mode in IDEs and apps.
2. TypeScript project setup
Previous conversation about setting up a TypeScript project with Bun.
`
When enabled, important conversation exchanges are automatically stored in Kybernesis Brain:
- Filters out greetings and simple commands
- Combines user question with assistant response for context
- Tags with configured capture tags (default: openclaw`)
- Generates concise titles from user messages
| Feature | Kybernesis | Supermemory |
|---------|------------|-------------|
| Search | Hybrid (vector + metadata + relationships) | Vector only |
| Tiering | Auto hot/warm/archive | Flat storage |
| Background Processing | Sleep agent (auto-tag, link, summarize) | None |
| Relationship Graph | Automatic entity linking | None |
| Connectors | Google Drive, Notion | None |
| Self-Hosted | Yes | Cloud only |
Get your API key from the Kybernesis Dashboard.
- Documentation
- GitHub Issues
- Discord
MIT