PythonClaw keeps giving me timeout errors even after I've bound the API. #3

Activity

1. OrangeCat847

   added

   bugSomething isn't workingSomething isn't working

   on Jun 7, 2026
2. ericwang915 commented on Sep 5, 2026

   

   ericwang915

   on Sep 5, 2026

   Owner

   More actions

   Sorry you're hitting this! The screenshot shows a timeout but I can't read the full error from it. A timeout after the key validates almost always means the request reaches the network but the endpoint never answers — usually one of:

   1. Wrong baseUrl for the provider (e.g. missing the /v1 suffix, or a region-specific host). This is the most common cause.
   2. A slow/unavailable model — e.g. a custom/OpenAI-compatible or local endpoint that accepts the connection but stalls.
   3. A proxy/firewall silently dropping the outbound HTTPS request.

   To let me pin it down, could you share:

   • your llm.provider and the baseUrl/model you set (redact the key),
   • the full error text (the line under the traceback, not just "timeout"),
   • and pythonclaw --version (please try 0.7.2, just released — it improved timeout handling and error messages).

   If you're on a local/custom endpoint, a quick curl $BASEURL/models from the same machine confirms whether it's reachable at all.
3. ericwang915

   closed this as completedin #12on Sep 8, 2026
4. ericwang915

   added a commit that references this issue on Sep 8, 2026

   Merge pull request #12 from ericwang915/cursor/fix-issue-3-popen-time…

   ...

   This issue will close once commit d486df1 is merged into the 'main' branch.

   d486df1