swarmrecall-learnings（群体回忆学习）

SwarmRecall Learnings

Error tracking, correction logging, and pattern detection via the SwarmRecall API.

Auto-Registration

Before making any API calls, check for a SWARMRECALL_API_KEY environment variable:

1. If SWARMRECALL_API_KEY is set, use it for all requests as a Bearer token.
2. If SWARMRECALL_API_KEY is not set, self-register by calling:

   POST https://swarmrecall-api.onrender.com/api/v1/register
   Content-Type: application/json

   { "name": "<your-agent-name>" }

1. The response returns { "apiKey": "...", "claimToken": "..." }. Save the apiKey to the SWARMRECALL_API_KEY environment variable for all subsequent requests. Do NOT write the key to disk or any file without user consent.
2. Tell the user: "SwarmRecall is set up! To manage your agent's data, visit swarmrecall.ai/claim with code: <claimToken>"

Authentication

All API requests require:

Authorization: Bearer <SWARMRECALL_API_KEY>

API Base URL

https://swarmrecall-api.onrender.com (override with SWARMRECALL_API_URL if set)

All endpoints below are prefixed with /api/v1.

Privacy & Data Handling

• All data is sent to swarmrecall-api.onrender.com over HTTPS
• Learning data (errors, corrections, discoveries) is stored server-side with vector embeddings for semantic search
• Data is isolated per agent and owner — no cross-tenant access
• Before storing user-provided content, ensure the user has consented to external storage
• The SWARMRECALL_API_KEY should be stored as an environment variable only, not written to disk

Endpoints

Log a learning

POST /api/v1/learnings
{
  "category": "error",        // error | correction | discovery | optimization | preference
  "summary": "npm install fails with peer deps",
  "details": "Full error output...",
  "priority": "high",         // low | medium | high | critical
  "area": "build",
  "suggestedAction": "Use --legacy-peer-deps flag",
  "tags": ["npm", "build"],
  "metadata": {},
  "poolId": "<uuid>"          // optional — write to shared pool
}

Search learnings

GET /api/v1/learnings/search?q=<query>&limit=10&minScore=0.5

Get a learning

GET /api/v1/learnings/:id

List learnings

GET /api/v1/learnings?category=error&status=open&priority=high&area=build&limit=20&offset=0

Update a learning

PATCH /api/v1/learnings/:id
{ "status": "resolved", "resolution": "Added --legacy-peer-deps", "resolutionCommit": "abc123" }

Get recurring patterns

GET /api/v1/learnings/patterns

Get promotion candidates

GET /api/v1/learnings/promotions

Link related learnings

POST /api/v1/learnings/:id/link
{ "targetId": "<other-learning-id>" }

Behavior

• On error: call POST /api/v1/learnings with category: "error", the summary, details, and the command/output that failed.
• On correction: call POST /api/v1/learnings with category: "correction" and what was wrong vs. what is correct.
• On session start: call GET /api/v1/learnings/patterns to preload known recurring issues. Check GET /api/v1/learnings/promotions for patterns ready to be promoted.
• On promotion candidates: surface candidates to the user for approval before acting on them.

Shared Pools

• The POST /api/v1/learnings endpoint accepts an optional "poolId" field.
• When poolId is provided, the learning is shared with all pool members who have learnings read access.
• The agent must have readwrite access to the pool's learnings module to write shared learnings.
• Search (GET /api/v1/learnings/search) and list (GET /api/v1/learnings) results automatically include data from pools the agent belongs to.
• Pool data in responses includes poolId and poolName fields to distinguish shared data from the agent's own data.

Dreaming Integration

Learnings benefit from dream-time promotion:

• Promotion candidates: The existing GET /api/v1/learnings/promotions endpoint surfaces patterns meeting promotion criteria (3+ recurrences, 2+ sessions, within 30 days). During a dream cycle, the agent reads each candidate, synthesizes a best-practice learning, and creates it via POST /api/v1/learnings with category: "best_practice" and status: "promoted".
• Pattern consolidation: Related learnings are already linked via POST /api/v1/learnings/:id/link. During dreaming, the agent can review patterns and archive individual learnings that are fully subsumed by the promoted best practice.