Repository navigation
ref(agents): Simplify OpenAI provider for gateway compatibility - #724
Merged
Merged
Conversation
dcramer
force-pushed
the
feat/openai-compatible-provider
branch
2 times, most recently
from
January 19, 2026 18:35
e36d041 to
972afb8
Compare
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
Remove reasoningEffort parameter from OpenAI provider - this was only applicable to o1/o3 reasoning models and could cause issues with OpenAI-compatible gateways that don't support it. Keep the critical settings that ensure compatibility: - structuredOutputs: false (required for nullable fields, fixes #623) - temperature: 1 (required by GPT-5, AI SDK defaults to 0) Add integration test that verifies the provider configuration works correctly with real OpenAI API calls (skipped when OPENAI_API_KEY not set). Co-Authored-By: Claude <noreply@anthropic.com>
dcramer
force-pushed
the
feat/openai-compatible-provider
branch
from
January 19, 2026 18:43
972afb8 to
4b013c6
Compare
Contributor
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.
Addresses bugbot feedback: the integration test only checked for OPENAI_API_KEY but didn't explicitly configure the OpenAI provider. Since auto-detect prefers Anthropic when both keys are set, the test could inadvertently use the wrong provider. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Remove
reasoningEffortparameter from OpenAI provider - this was onlyapplicable to o1/o3 reasoning models and could cause issues with
OpenAI-compatible gateways that don't support it.
Kept the critical settings that ensure compatibility:
structuredOutputs: false(required for nullable fields, fixes 0.20.0 fails to communicate to the api due to missing query (self hosted) #623)temperature: 1(required by GPT-5, AI SDK defaults to 0)Added integration test that actually calls
searchIssuesAgentwith realOpenAI API (unlike unit tests which mock AI SDK, or evals which don't execute
tool handlers). The test:
OPENAI_API_KEYnot set (safe for CI without secrets)This fills a testing gap:
Usage with gateways:
sentry-mcp --openai-base-url=https://gateway.ai.vercel.com/v1 OPENAI_MODEL=openai/gpt-4o # if gateway expects different format