Migrate from Self-Hosted OpenClaw

Moving from self-hosted OpenClaw to Curate-Me takes about 5 minutes. Your OpenClaw config stays the same. Your agents don’t change. You just point at a different URL and get governance for free.

Before vs After

# BEFORE: Self-hosted OpenClaw
# You manage: VPS, Docker, security patching, cost monitoring
OPENAI_BASE_URL=https://api.openai.com/v1
# Your agents call OpenAI directly — no governance, no cost control
 
# AFTER: Curate-Me managed
# We manage: hosting, security, cost governance, audit trails
OPENAI_BASE_URL=https://api.curate-me.ai/v1/openai
X-CM-API-Key=cm_sk_xxx
# Your agents go through our governance chain — same results, full control

Step-by-Step Migration

1. Create Your Account

# Sign up at dashboard.curate-me.ai
# Get your API key: cm_sk_xxx
# No credit card required

2. Update Environment Variables

Replace your LLM provider base URLs:

Add the Curate-Me API key header to all requests:

X-CM-API-Key: cm_sk_xxx

3. Configure Governance Policies

From the dashboard (dashboard.curate-me.ai):

1. Cost controls: Set daily budget (e.g., $25/day)
2. PII scanning: Toggle ON (enabled by default on paid plans)
3. Rate limiting: Set per-key request limits
4. Model allowlist: Restrict which models your org can use
5. HITL approvals: Enable for operations above a cost threshold

4. Verify

# Test health
curl https://api.curate-me.ai/v1/health \
  -H "X-CM-API-Key: cm_sk_xxx"
 
# Test a completion
curl https://api.curate-me.ai/v1/openai/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-CM-API-Key: cm_sk_xxx" \
  -H "Authorization: Bearer sk-your-openai-key" \
  -d '{"model": "gpt-4o", "messages": [{"role": "user", "content": "Hello"}]}'

What Stays the Same

• Your OpenClaw configuration (openclaw.json)
• Your LLM provider API keys
• Your agent code
• Your skills and templates
• Your messaging channel integrations

What Changes

Cost Comparison

Common Questions

Q: Will my agents notice the change? No. The gateway proxies requests transparently. Your agents receive the same responses from the same LLM providers. The governance chain runs before the request is forwarded.

Q: What about latency? The governance chain adds ~5-15ms per request (policy evaluation + PII scan). For most use cases this is imperceptible.

Q: Can I use multiple providers? Yes. The gateway supports 17+ providers including OpenAI, Anthropic, Google, Groq, Mistral, xAI, and more simultaneously. Each request is routed to the correct provider based on the URL path.

Q: What if I want to go back to self-hosted? Change the URL back. No lock-in. Your data is exportable from the dashboard.

Next Steps

• OpenClaw Quickstart
• Security Best Practices
• API Reference