OpenClaw / BetterClaw
Autonomous health agents on OpenClaw, with Betterness as the data backbone.
What is BetterClaw
BetterClaw is Betterness's extension for OpenClaw agents — starter kits, agent templates, and MCP wiring for health agent workflows. If you build a health agent from scratch, you'd need to solve consent flows, wearable integrations and lab pipelines yourself. BetterClaw provides that infrastructure.
Architecture
- OpenClaw runtime — agent execution environment
- Betterness MCP — governed health tool access
- Your LLM — Claude, GPT, Gemini, or any model
Quick start
- Create an API key at /console/keys. Narrow it to the permissions the agent needs.
- Wire MCP into your agent:
{
"mcpServers": {
"betterness": {
"url": "https://api.betterness.ai/mcp",
"headers": {
"Authorization": "Bearer bk_your_api_key_here"
}
}
}
}
Agent patterns
Daily briefing agent
Call get_updates for what changed since the last run, then getActivityData, getSleepData, getVitals, searchBiomarkers and getBiologicalAge. Pass the previous cursor to get_updates on each run.
Lab gap analysis agent
searchBiomarkers— recent resultsgetLoincCodes— LOINC slug lookup- Detect gaps (no recent Vitamin D, thyroid, lipid panel, etc.)
listAvailableLabTests(confirmedUsState, loincSlug)— find tests covering gaps- Recommend with name, price, included biomarkers
Proactive monitoring agent
searchBiomarkers(range: "OUT_OF_RANGE")- Compare against previous values for trends
- Alert with actionable recommendations
Governance
- Keys can be narrowed to some permissions and carry an expiry date
- Calls are rate-limited per member — see Limits & scopes
- Every tool call is recorded in the audit trail
- A purchase, lab order submission, appointment change or message is only prepared until the agent calls
confirm_actionwith itspendingId: build the member's approval into that step. A key withoutpurchases,labs:orderorinbox:writecannot prepare them at all.

