Skip to content

Troubleshooting ​

Server not responding

Confirm the server is running and the port matches .llmockrc.json. Open http://localhost:8001 — if it's unreachable, the server may not have started.

llmock start exits with an error

Run llmock start --foreground with the same options to see the server's output. If it reports a server already running on the port, run llmock stop (with --port if it isn't 8001) first.

The server refuses to start on a malformed responseRules entry, or on a storedResponsesFile that is missing or not valid; the message names the setting and the file. It also refuses a chaos setting that is not valid (see Chaos mode).

Calls fail with a 500 or another error you didn't expect

Check whether chaos is on: the Chaos box on the dashboard's Settings page shows ENABLED, and a chaos error carries the header x-llmock-chaos: true. Set "chaos": { "enabled": false } in the preset (or drop --chaos=true) and restart.

Port already in use

Change the port in .llmockrc.json or pass it as a flag:

bash
llmock start --port=8002

Request validation failures

  • For the claude preset, model, max_tokens and messages are all required
  • Confirm your request template matches the provider's API format
  • Check the request shape with View request log on the dashboard
  • Verify the name field in your model config matches the template filename prefix

Response delays not applied

Ensure responseDelay.min or responseDelay.max is greater than 0, then restart the server.

A fixture or stored reply returns an error

The file could not be read when the request arrived. Check that the path is relative to the folder holding .llmockrc.json and, in Docker, that the file is mounted into the container.

Released under the MIT License.