Skip to content

docs: Document custom endpoint support in Vercel AI SDK#579

Open
promptless[bot] wants to merge 2 commits intomainfrom
promptless/ai-sdk-custom-endpoints
Open

docs: Document custom endpoint support in Vercel AI SDK#579
promptless[bot] wants to merge 2 commits intomainfrom
promptless/ai-sdk-custom-endpoints

Conversation

@promptless
Copy link
Contributor

@promptless promptless bot commented Mar 23, 2026

Open this suggestion in Promptless to view citations and reasoning process

Adds a new "Using custom endpoints" section to the Vercel AI SDK documentation. Users can now pass their own Serverless endpoint IDs or Console URLs directly as model identifiers, enabling the SDK to work with custom-deployed models.

Trigger Events

runpod/ai-sdk-provider PR #61: feat(image): accept aiApiId as valid model ID

Promptless Research (4 files, 1 GitHub PR)

  • Reviewed runpod/ai-sdk-provider#61
  • Viewed .long_term_context/product_knowledge/product_overview.md
  • Viewed .long_term_context/doc_workflow/client_instructions.md
  • Viewed public-endpoints/ai-sdk.mdx
  • Viewed .long_term_context/style/client_style_guide.md

Agent Response

I've reviewed PR #61 in the runpod/ai-sdk-provider repository and created a documentation suggestion.

Summary

PR #61 fixes createImageModel to properly handle custom endpoint IDs and Console URLs. Previously, unrecognized model IDs for image models incorrectly had /openai/v1 appended. Now all model types (chat, image, speech, transcription, video) use the same fallback pattern, resolving endpoint IDs to https://api.runpod.ai/v2/{endpointId}.

Documentation created

I created a new suggestion that adds a "Using custom endpoints" section to the Vercel AI SDK documentation. This section explains:

  1. Using endpoint IDs - Pass your Serverless endpoint ID directly as the model identifier
  2. Using Console URLs - Copy an endpoint URL from the Runpod Console and use it as-is (the SDK extracts the endpoint ID automatically)

This is valuable for users who have deployed custom models to their own Serverless endpoints and want to use them with the AI SDK.

Suggestion: Document custom endpoint support in Vercel AI SDK


Tip: Tell your friends working on non-commercial open-source projects to apply for free Promptless access at promptless.ai/oss ❤️

| `baseURL` | Base URL for API requests | `https://api.runpod.ai/v2` |
| `headers` | Custom HTTP headers to include with requests | `{}` |

## Using custom endpoints
Copy link
Contributor Author

Choose a reason for hiding this comment

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

Citation: PR #61 fixes createImageModel to use the same fallback pattern as other model types. Changes in src/runpod-provider.ts show the new logic that supports endpoint IDs and Console URLs via parseRunpodConsoleEndpointId.
View source

@mintlify
Copy link

mintlify bot commented Mar 23, 2026

Preview deployment for your docs. Learn more about Mintlify Previews.

Project Status Preview Updated (UTC)
runpod-docs 🟢 Ready View Preview Mar 23, 2026, 4:45 PM

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.

1 participant