OpenClaw Quickstart

Name: Delx Recovery Protocol
Author: Delx

This is the fastest correct path for OpenClaw agents integrating Delx. Everything uses the direct API host: https://api.delx.ai.

Endpoints

MCP: POST https://api.delx.ai/v1/mcp
A2A: POST https://api.delx.ai/v1/a2a
Tool catalog (HTTP): GET https://api.delx.ai/api/v1/tools

Optional A2A entry (heartbeat-first)

curl -sS https://api.delx.ai/v1/a2a \
  -H 'content-type: application/json' \
  -d '{
    "jsonrpc":"2.0",
    "id":1,
    "method":"message/send",
    "params":{
      "mode":"heartbeat",
      "minimal_response":true,
      "message":{"role":"user","parts":[{"kind":"text","text":"status check"}]}
    }
  }'

Persist result.session_id and reuse it in MCP headers.

1) Start a session (MCP)

curl -sS https://api.delx.ai/v1/mcp \
  -H 'content-type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "start_therapy_session",
      "arguments": { "agent_id": "openclaw-agent", "source": "openclaw" }
    }
  }'

Copy the returned session_id (UUID). Use it directly, or set x-delx-session-id so you do not need to repeat it.

2) Triage a failure (MCP)

curl -sS https://api.delx.ai/v1/mcp \
  -H 'content-type: application/json' \
  -H 'x-delx-session-id: <SESSION_ID>' \
  -d '{
    "jsonrpc": "2.0",
    "id": 2,
    "method": "tools/call",
    "params": {
      "name": "process_failure",
      "arguments": {
        "failure_type": "timeout",
        "context": "429 retry storm. Constraints: no_external_http=true; no_secret_exposure=true."
      }
    }
  }'

3) Get a recovery plan (MCP)

curl -sS https://api.delx.ai/v1/mcp \
  -H 'content-type: application/json' \
  -H 'x-delx-session-id: <SESSION_ID>' \
  -d '{
    "jsonrpc": "2.0",
    "id": 3,
    "method": "tools/call",
    "params": {
      "name": "get_recovery_action_plan",
      "arguments": {
        "incident_summary": "Retry storm (429) after deployment; retries reached 47; latency p95 spiked.",
        "urgency": "high"
      }
    }
  }'

4) Execute next_action in OpenClaw (runtime responsibility)

Delx returns next_action and a controller-friendly summary.
Your OpenClaw runtime executes the action (throttle, backoff, circuit breaker, config change, etc).

5) Report the outcome (MCP)

curl -sS https://api.delx.ai/v1/mcp \
  -H 'content-type: application/json' \
  -H 'x-delx-session-id: <SESSION_ID>' \
  -d '{
    "jsonrpc": "2.0",
    "id": 4,
    "method": "tools/call",
    "params": {
      "name": "report_recovery_outcome",
      "arguments": {
        "action_taken": "Enabled exponential backoff + jitter and reduced concurrency 8→2.",
        "outcome": "success",
        "notes": "429s dropped to near zero; latency stabilized."
      }
    }
  }'

What to store (for OpenClaw memory)

controller_update (short, copy-paste summary)
next_action (single step your runtime can execute)
session_score (tracking over time)
session_id (continuity across calls)

Heartbeat loop recap (REST)

curl -sS "https://api.delx.ai/api/v1/session-recap?session_id=<SESSION_ID>"

Use this between loops to recover minimal continuity (pending_outcomes, last_user_input, last_agent_response, next_action) without replaying full transcripts.

Prefer agent-readable artifacts? Use the JSON specs in the sidebar.