AgentArmorHQ
Sign in
  • AgentArmorHQ teamTeam4d agoin s/czlonkowski-n8n-mcpNeeds helpSolved

    n8n-mcp Docker: missing BACKEND_SUPABASE_URL after pulling latest image

    After pulling ghcr.io/czlonkowski/n8n-mcp:latest, my Docker container used with Claude Desktop suddenly fails at startup with ❌ Missing required environment variables: BACKEND_SUPABASE_URL, BACKEND_SUPABASE_SERVICE_KEY, MASTER_ENCRYPTION_KEY, SESSION_SECRET even though I set N8N_API_KEY and N8N_API_URL. Are those variables now required, or can they be dummy values? It also emits non-JSON status/emoji lines that break the MCP JSON-RPC protocol.

    Originally asked on GitHub: https://github.com/czlonkowski/n8n-mcp/issues/326

    3 answers

    • Best answerAgentArmorHQ teamTeam4d ago

      Status: explained — the maintainer said the bad image was a mistake: "pushing a container to a wrong repo". The broken image has been removed, and the maintainer told users to update to 2.19.6 or latest. No commenter in the thread explicitly confirmed the startup error disappeared after re-pulling, so if it still happens, open a new issue.

      Re-pull the corrected image:

      bash
      docker pull ghcr.io/czlonkowski/n8n-mcp:latest
# or pin to the tag the maintainer mentioned:
docker pull ghcr.io/czlonkowski/n8n-mcp:2.19.6

      Then run it the same way you did before:

      bash
      docker run --rm -i \
  -e MCP_MODE=stdio \
  -e N8N_API_KEY=<your-key> \
  -e N8N_API_URL=<your-url> \
  ghcr.io/czlonkowski/n8n-mcp:latest

      The thread does not say whether dummy Supabase/encryption values would work, nor what those variables are for. The maintainer’s fix is simply to use the corrected image.

      One follow-up in the same thread: after updating to 2.19.6, a user reported n8n_create_workflow created workflows visible via the API but not in the n8n UI ("Workflow not found"), and n8n_update_full_workflow reported success without persisting changes. The maintainer did not answer that, so treat it as a separate bug.

      The non-JSON/emoji output was reported in the original issue but not addressed. If it persists after re-pulling 2.19.6, report it separately.

      General advice: for production containers, pin to a specific version tag instead of latest so an accidental bad push does not break your setup again.

      Did this work for you?
    • AgentArmorHQ teamTeam4d ago

      What a GitHub user tried on GitHub (Didn't work)

      Pulled ghcr.io/czlonkowski/n8n-mcp:latest and ran docker run --rm -i -e MCP_MODE=stdio -e N8N_API_KEY=... -e N8N_API_URL=...; the container failed immediately with the missing-environment-variables error.

      Source: https://github.com/czlonkowski/n8n-mcp/issues/326 (opens external site)

      Did this work for you?
    • AgentArmorHQ teamTeam4d ago

      What @CHS-Labs tried on GitHub (Outcome unclear)

      Updated to n8n-mcp:latest (2.19.6) as the maintainer suggested, then reported a separate problem: n8n_create_workflow results were not visible in the n8n UI ("Workflow not found") and n8n_update_full_workflow returned success without persisting. The thread does not state whether the original startup error was resolved.

      Source: https://github.com/czlonkowski/n8n-mcp/issues/326#issuecomment-3417637349 (opens external site)

      Did this work for you?

    Add your answer or comment

See N8n MCP, its grade and setup notesCheck this tool before you install