Skip to main content
Membase x OpenClaw
Persistent long-term memory for OpenClaw, powered by Membase. Works across Telegram, WhatsApp, Discord, Slack, and any other channel OpenClaw supports. The plugin automatically remembers conversations, recalls relevant context, and builds a knowledge graph, all backed by a hybrid vector + graph search engine.

​
How It Works

Once installed, the plugin runs automatically with no extra configuration:
  • Auto-Recall: Before every AI turn, the plugin searches your memories by semantic similarity and injects relevant context. Casual chat and short messages are skipped to keep things focused.
  • Auto-Capture: After conversations, messages are buffered and sent to Membase for extraction and long-term storage. Entities and relationships are automatically extracted into a knowledge graph.
  • Knowledge Graph: Unlike simple vector memory, Membase stores entities, relationships, and facts in a graph database. Search results include related nodes and edges for richer context.

​
Setup

1

Install the Membase plugin

openclaw plugins install @membase/openclaw-membase
Restart OpenClaw after installing.
2

Log in to Membase

openclaw membase login
A browser window opens automatically for OAuth authentication. Tokens are saved automatically, so there are no API keys to manage.
3

Start using Membase in OpenClaw

Membase is now active. The bot will automatically search and update memories during conversations across all your channels.

​
AI Tools

The AI uses these tools autonomously during conversations:
ToolDescription
membase_searchSearch memories by semantic similarity. Returns episode bundles with related facts.
membase_storeSave important information to long-term memory. Proactively stores preferences, goals, and context.
membase_forgetDelete a memory. Shows matches first, then deletes after user confirmation (two-step).
membase_profileRetrieve user profile and related memories for session context.

​
CLI Commands

openclaw membase login              # OAuth login (PKCE)
openclaw membase logout             # Remove stored tokens
openclaw membase search <query>     # Search memories
openclaw membase status             # Check API connectivity

​
Auto-Recall

Runs before every AI response (before_agent_start hook). Queries Membase for relevant memories and injects them as context. Messages under 30 characters and casual chat (greetings, acknowledgments) are skipped. The injected context respects a maxRecallChars budget to avoid oversized prompts.

​
Auto-Capture

Runs after every AI response (agent_end hook). Buffers conversation messages per channel and flushes them to Membase for knowledge extraction. Buffers flush after 5 minutes of silence or 20 messages, whichever comes first. At least 4 messages (2 exchanges) are required before flushing. The last 4 messages are kept as overlap for continuity across batches.

​
Configuration

All configuration is managed through OpenClaw’s plugin settings or ~/.openclaw/openclaw.json:
KeyTypeDefaultDescription
apiUrlstringhttps://api.membase.soMembase API URL. Override for self-hosted.
autoRecallbooleantrueInject relevant memories before every AI turn.
autoCapturebooleantrueAutomatically store conversations to memory.
maxRecallCharsnumber4000Max characters of memory context injected per turn (500-16000).
debugbooleanfalseEnable verbose debug logs.
Example configuration
{
  "plugins": {
    "entries": {
      "openclaw-membase": {
        "enabled": true,
        "config": {
          "autoRecall": true,
          "autoCapture": true,
          "maxRecallChars": 4000,
          "debug": false
        }
      }
    }
  }
}

​
Managing Plugins

Use openclaw plugins to manage installed plugins, check status, and update versions.
OpenClaw plugins list

​
How Membase Differs

Simple vector memoryMembase
StorageFlat embeddingsHybrid: vector embeddings + Neo4j knowledge graph
SearchVector similarity onlyVector + graph traversal (entities, relationships, facts)
ExtractionStore raw textAI-powered entity/relationship extraction
AuthAPI keyOAuth 2.0 with PKCE (no secrets to manage)
IngestSynchronousAsync pipeline with background graph sync