n8n + PicoClaw for self-hosted AI webhooks

n8n is a popular workflow engine for self-hosters. Pair it with PicoClaw when you want an LLM step that stays easy to operate: n8n handles branching, retries, and integrations; PicoClaw handles the assistant logic with a minimal footprint.

1. Typical flow

  1. An n8n trigger fires (schedule, webhook, email, Slack, etc.).
  2. n8n formats the payload (JSON, text, template).
  3. An HTTP node calls PicoClaw or your gateway endpoint.
  4. The response is parsed and sent to the next destination (chat, database, ticket system).

2. Expose PicoClaw safely

Prefer private Docker networks, VPN, or reverse proxies with authentication. Never leave an open LLM webhook on the public internet without rate limits and auth. Review Security and your reverse-proxy docs before going live.

3. Payload design

Keep prompts short and structured. Store secrets in n8n credentials, not in workflow JSON exports. If a run can recurse, add guards so a failed node does not retry expensive LLM calls indefinitely.

4. When to use n8n vs cron

For simple schedules, cron + PicoClaw may be enough. Use n8n when you need visual debugging, many SaaS connectors, or complex branching.

5. Next steps