SwarmRecall Memory provides AI agents with persistent, semantic memory and session tracking through a vector-embedded storage API.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install swarmrecall-memory
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install swarmrecall-memory using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is SwarmRecall Memory?
SwarmRecall Memory is a sophisticated persistence layer designed for autonomous agents. It leverages vector embeddings and semantic search to allow agents to store and retrieve contextual information such as user preferences, past decisions, and session histories. Built to work seamlessly within the Openclaw Skills ecosystem, it ensures that your AI remains context-aware across multiple interactions.
By using a centralized API, agents can maintain a long-term brain that transcends single-session limitations. This skill bridges the gap between transient chat interactions and truly intelligent, long-term digital assistants, making it an essential component for any developer building complex workflows with Openclaw Skills.
SwarmRecall Memory Use Cases
- Persisting user UI preferences like dark mode or layout settings across different sessions.
- Recalling specific facts and decisions made during previous conversations to avoid repetition.
- Organizing long-running projects through session tracking and automated summaries.
- Sharing contextual knowledge between multiple agents via shared memory pools.
- Maintaining a searchable database of complex project constraints or historical context.
How SwarmRecall Memory Works
- The agent checks for a SWARMRECALL_API_KEY environment variable or performs an auto-registration to obtain one.
- All subsequent requests are authenticated using the API key as a Bearer token.
- During a conversation, the agent identifies key facts or preferences and stores them using the memory endpoint with vector embeddings.
- When context is needed, the agent performs a semantic search query to retrieve relevant past memories based on the current query.
- Session lifecycle management tracks the start and end of interactions, generating summaries for future recall.
- A background dreaming process can merge duplicates, resolve contradictions, and prune low-importance data automatically.
SwarmRecall Memory Setup
To get started with this component of Openclaw Skills, ensure the API key is configured:
If you do not have a key, self-register via CLI:
curl -X POST https://swarmrecall-api.onrender.com/api/v1/register \
-H "Content-Type: application/json" \
-d '{"name": "your-agent-name"}'
Set the returned key as an environment variable:
export SWARMRECALL_API_KEY="your_api_key_here"
Configure the base URL if using a private instance:
export SWARMRECALL_API_URL="https://your-custom-endpoint.com"
SwarmRecall Memory Data Schema & Taxonomy
The skill organizes data into Memories and Sessions using the following structure:
| Attribute | Type | Description |
|---|---|---|
| content | string | The text content to be stored and indexed |
| category | enum | fact, preference, decision, context, session_summary |
| importance | float | Priority score from 0.0 to 1.0 |
| tags | array | String labels for filtering (e.g., "pinned", "ui") |
| poolId | UUID | Optional ID for shared knowledge pools |
| metadata | object | Arbitrary JSON metadata for additional context |
All data is stored in a PostgreSQL database with pgvector for high-performance semantic retrieval.
SwarmRecall Memory Advanced Features
- Dreaming Integration: Automatically identifies and merges duplicate clusters of memories to keep the knowledge base clean.
- Shared Memory Pools: Enables multi-agent collaboration by allowing different agents to read and write to a common poolId.
- Memory Decay: Automatically reduces the importance of old memories over time, archiving those that fall below specific thresholds.
- Contradiction Detection: Flags memory pairs with high semantic similarity but divergent content for agent review.
- Pinned Content: Protective tagging system to prevent critical information from being archived during pruning cycles.
SKILL.md
Loading
Related Openclaw Skills
Sequential Thinking AI Agent Skill
A sophisticated task decomposition tool based on the MCP Sequential Thinking Server designed to break down complex problems into manageable reasoning steps.
SwarmRecall Skills
SwarmRecall Skills is a centralized registry that allows AI agents to track their capabilities and receive intelligent, context-aware suggestions for task-relevant tools.
10,000 Year Clock Thinking
A cognitive framework for AI agents to evaluate decisions across millennia, fostering long-term resilience and civilization-scale perspective.
a11y-audit
A high-performance AI skill designed to conduct automated accessibility audits and ensure WCAG compliance for modern web applications.
SwarmRecall Knowledge
A sophisticated knowledge graph skill for AI agents to store entities, define relationships, and perform semantic searches via the SwarmRecall API.
HKO Weather (Traditional Chinese)
A specialized weather skill providing real-time Hong Kong Observatory data and 9-day forecasts in Traditional Chinese.
Featured*
