"failed to discover Ollama models: TypeError: fetch failed"
This means OpenClaw's startup auto-discovery tried to query Ollama for available models and the request failed, usually because Ollama isn't running, hasn't finished loading a model yet, or the baseUrl is unreachable. Pre-load your model so it's ready when discovery runs, verify the baseUrl is reachable from where OpenClaw runs, then restart OpenClaw to re-run discovery.
Likely cause
OpenClaw tried to list Ollama models at startup but the request failed - usually because Ollama is not running, has not finished loading a model yet, or the baseUrl is unreachable.
The fix
- 1 Start Ollama and pre-load the model so it is ready when discovery runs.
- 2 Verify the baseUrl is reachable from where OpenClaw runs.
- 3 Restart OpenClaw so model discovery re-runs.
curl http://127.0.0.1:11434/api/tagsHit a different error?
Paste any agent error and get the cause and fix in seconds.
Frequently asked questions
How is this different from a plain 'fetch failed'?
Plain 'fetch failed' means the HTTP connection itself failed. 'Failed to discover models' means the connection may partially work but the model-list query specifically fails, usually because Ollama hasn't finished loading a model yet.
How do I stop discovery from racing the model load?
Pre-load the model first (for example ollama run
It still times out even after the model is loaded. Why?
Large models can take 30-60 seconds to load and may exceed the discovery timeout, and low free RAM slows everything down. Ensure enough memory (16GB+ for a 7B model) and pre-load before startup.
Stop firefighting agent errors
Decoding errors one at a time is the manual version of what BetterClaw automates. Run your OpenClaw agents hosted with managed models, retries and config validation built in.
$19/month per agent · BYOK · 7-day money-back guarantee