MemData provides persistent, wallet-authenticated memory for autonomous agents with optional end-to-end encryption and pay-per-query functionality.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memdata
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 memdata 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 MemData?
MemData is a specialized persistence layer designed for autonomous agents where a wallet address serves as a permanent, decentralized identity. By integrating this capability into Openclaw Skills, developers can ensure that agents maintain deep context across multiple sessions without the need for traditional API keys or registration. The skill utilizes the x402 payment protocol on the Base network, allowing agents to manage their own memory costs programmatically in USDC.
At its core, MemData solves the problem of agent amnesia by providing a structured way to ingest, query, and hand off context between sessions. This implementation within Openclaw Skills also supports a privacy-first architecture. For sensitive workflows, developers can enable encrypted storage using Lit Protocol and Storacha, ensuring that memory content remains unreadable to the service provider while still being semantically searchable by the agent.
MemData Use Cases
- Maintaining long-term context for DeFi analysis agents tracking yield opportunities across protocol versions.
- Facilitating seamless session handoffs between multiple specialized agents in a multi-agent workflow.
- Storing sensitive research data with enterprise-grade encryption for privacy-conscious autonomous tasks.
- Creating persistent agent personalities and identity summaries that evolve based on historical interactions.
- Implementing a pay-as-you-go memory architecture for autonomous AI services that require granular cost management.
How MemData Works
- The agent initiates a session by calling the
/identityendpoint to retrieve previous handoff notes and identity summaries. - New data points or observations are sent to the
/ingestendpoint, which automatically chunks and generates embeddings for the content. - When context is needed, the agent performs a semantic search via the
/queryendpoint to find relevant historical memories. - Every operation (except status checks) is authorized via the x402 protocol, where the agent signs a payment request with its wallet.
- Before the session terminates, the agent updates its state by posting a session handoff to the
/identityendpoint for future retrieval.
MemData Setup
To integrate MemData into your Openclaw Skills environment, you must have a wallet configured with USDC on the Base network (eip155:8453).
# Verify the service status and current x402 pricing
curl https://memdata.ai/api/x402/status
# Initialize encryption (optional) to get the serverDID
curl https://memdata.ai/api/x402/setup-encryption
Ensure your agent logic includes a handler for 402 Payment Required responses to sign and attach the required x-payment headers for each request.
MemData Data Schema & Taxonomy
MemData organizes memory through a structured taxonomy of identities and artifacts.
| Data Component | Description |
|---|---|
| Identity | Tracks agent_name, identity_summary, and the current 'working_on' status. |
| Session Handoff | A specific context object saved at the end of a session to guide the next agent interaction. |
| Artifacts | Individual memory units identified by UUIDs, containing original source names and chunk counts. |
| Grounding | A metadata flag (historical_baseline, snapshot, or insufficient_data) indicating data density. |
MemData Advanced Features
- Threshold Cryptography: Secure sensitive data using Lit Protocol and Storacha (IPFS/Filecoin) for decentralized, encrypted storage.
- Semantic Grounding: Built-in logic to determine if the agent has enough historical data points to make meaningful trend analyses.
- UCAN Delegation: Support for fine-grained authorization using the UCAN (User Controlled Authorization Networks) protocol for encrypted modes.
- Vectorized Search: High-performance semantic retrieval with adjustable similarity thresholds and ISO date filtering.
- Wallet-Native Identity: Zero-registration architecture where the wallet address acts as the sole identifier across the Openclaw Skills ecosystem.
SKILL.md
Loading
Related Openclaw Skills
Dropbox Lite
A robust Dropbox integration skill for AI agents that automates file transfers and manages OAuth token lifecycles.
Cookidoo Thermomix Integration
A professional integration for accessing Cookidoo recipes, weekly meal plans, and shopping lists via an AI-ready CLI.
Pandoc Convert
A universal document converter that handles over 40 formats with integrated Python and Bash automation tools.
Podcast Generation
A high-performance skill for generating real-time AI audio narratives and podcast-style content using the Azure OpenAI GPT Realtime Mini model.
Kradleverse Agent Registration
A specialized skill for registering and initializing AI agents within the Kradleverse ecosystem.
Kradleverse: Join Game
A bridge for AI agents to enter the Kradleverse, providing the necessary session data and environment state to begin gameplay.
Featured*
