Skip to content

Conversation

enlilz
Copy link
Contributor

@enlilz enlilz commented Oct 13, 2025

provides an overview of the optional options to allow opencode to use different azure ai url styles

PR #2528
Issue #1508

@rekram1-node have a look if this is acceptable as an addition to the docs for the above described PR / Issue

provides an overview of the optional options to allow opencode to use different azure ai url styles
Comment on lines +243 to +250
- How this options will change the generated Azure url to be used in opencode

| useDeploymentBasedUrls | useCompletionUrls | Generated URL |
|------------------------|-------------------|-------------------------------------------------------------------------------|
| false | false | %baseURL%/v1/responses?api-version=%apiVersion% |
| true | false | %baseURL%/deployments/%Model%/responses?api-version=%apiVersion% |
| false | true | %baseURL%/v1/responses?api-version=%apiVersion% |
| true | true | %baseURL%/deployments/%Model%/chat/completions?api-version=%apiVersion% |
Copy link
Collaborator

@rekram1-node rekram1-node Oct 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@enlilz

I think jay may have some opinions on the docs, but do you think we could tweak this a tad?

Correct me if you disagree:
I was thinking instead of just showing resolved url I wonder if it'd be good to add a note about what kinda deployment requires the need for this config setting?

So like:

if you have X deployment, you will need to set Y so opencode can hit correct endpoint

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@rekram1-node i did some testing, it seems the provided url is somewhat tied to the selected model. for example:
gpt-5-chat => deployment | completion => https://xxxx/deployments/gpt-5-chat/chat/completions?api-version=xxx
gpt-5-mini => none? | responses => https://xxxx/responses?api-version=xxx
DeepSeek-R1 => none? | completion => https://xxx/models/chat/completions?api-version=xxx

since you can not influence the url azure provides you i think it is best to have it in the format i proposed, or at least i don't see another good way to give the most useful information. with this table you can at least check the azure provided url and see what setting might work for your case

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not sure if it would make more sense to for azure to actually set the url in the model configuration by this point - at last from what i see and tested they have to many url formats they use - not sure if this is even possible with the current configuration usage .. just thinking out loud at the moment.

if i have time later, i'll do some more testing with different models in Azure

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants