tinyhumansai/openhuman · error · Error

OpenAI Responses API timed out after ${timeoutMs}ms

Error message

OpenAI Responses API timed out after ${timeoutMs}ms

What it means

The fetch() to https://api.openai.com/v1/responses was aborted by the script's own AbortController after 300_000 ms (5 minutes); the resulting AbortError is caught and rethrown as this timeout error. Note the payload is already bounded to MAX_PROMPT_CHARS = 120_000, so this is wall-clock latency, not unbounded input.

Source

Thrown at scripts/release/generate-release-notes.mjs:492

  }

  const timeoutMs = 300_000;
  const controller = new AbortController();
  const timeout = setTimeout(() => controller.abort(), timeoutMs);
  let response;
  try {
    response = await fetch('https://api.openai.com/v1/responses', {
      method: 'POST',
      signal: controller.signal,
      headers: {
        Authorization: `Bearer ${key}`,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify(request),
    });
  } catch (error) {
    if (error?.name === 'AbortError') {
      throw new Error(`OpenAI Responses API timed out after ${timeoutMs}ms`);
    }
    throw error;
  } finally {
    clearTimeout(timeout);
  }

  const body = await response.text();
  let json = null;
  try {
    json = JSON.parse(body);
  } catch {
    // Keep the raw body in the error below.
  }

  if (!response.ok) {
    const message = json?.error?.message || body;
    throw new Error(`OpenAI Responses API failed (${response.status}): ${message}`);
  }

View on GitHub (pinned to a221052e0d)

Solutions

  1. Re-run — transient stalls are the most common cause
  2. Shrink the range: --from a more recent tag so fewer PRs are summarized
  3. Use --no-ai to render deterministic notes without calling OpenAI
  4. Check https://status.openai.com if it reproducibly times out on small ranges

Example fix

# before
node scripts/release/generate-release-notes.mjs --from v0.50.0   # 300s timeout over a huge range

# after
node scripts/release/generate-release-notes.mjs --from v0.58.0   # smaller window
# or
node scripts/release/generate-release-notes.mjs --from v0.50.0 --no-ai
Defensive patterns

Strategy: retry

Validate before calling

# bound the work before calling the API — huge ranges are the usual timeout cause
count=$(git rev-list --count "${FROM}..${TO}")
if [ "$count" -gt 300 ]; then
  echo "$count commits — using --no-ai" >&2
  exec node scripts/release/generate-release-notes.mjs --from "$FROM" --to "$TO" --no-ai
fi

Try / catch

for attempt in 1 2; do
  node scripts/release/generate-release-notes.mjs --from "$FROM" && break
  [ "$attempt" -eq 2 ] && { echo 'generation failed twice' >&2; exit 1; }
  echo 'retrying after timeout...' >&2; sleep 60
done

Prevention

When it happens

Trigger: A large release range producing a long generation that exceeds 5 minutes; a network stall between the runner and api.openai.com; OpenAI-side degradation on long completions.

Common situations: Big-bang merges producing hundreds of PRs in one range; CI runners with constrained or proxied egress; retrying during an OpenAI incident window.

Understand the failure class

Related errors

AI-assisted analysis of tinyhumansai/openhuman@a221052e0d (2026-08-16). Data as JSON: /api/errors/9af5522732511730. Report an issue: GitHub.