Version Packages

[github-actions]

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and publish to npm yourself or setup this action to publish automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@openrouter/ai-sdk-provider@1.3.0

Minor Changes

• #263 ed2bec5 Thanks @talos! - Add support for debug flag to echo upstream request body

  • Added debug option to OpenRouterChatSettings with echo_upstream_body boolean
  • The debug flag is passed through to the OpenRouter API in both streaming and non-streaming requests
  • Debug mode only works with streaming requests and returns the upstream request body (with sensitive data redacted) as the first chunk
  • Updated README with usage documentation

Patch Changes

• #266 c3f6381 Thanks @subtleGradient! - Fix token details in providerMetadata to only be included when present in API response. Previously, promptTokensDetails and completionTokensDetails were always included with default values of 0, which could be misleading. Now they are only included when the API actually returns these details, matching the behavior of costDetails.

• #268 6ac3814 Thanks @jamespsterling! - Add reasoning_details accumulation and providerMetadata support for multi-turn conversations

  • Accumulate reasoning_details from reasoning parts when converting messages
  • Include reasoning_details in providerMetadata for reasoning delta chunks during streaming
  • Enables users to accumulate reasoning_details across multi-turn conversations

• #246 78d20ef Thanks @abromberg! - build was failing due to mismatch with @openrouter/sdk types. simply renamed models.Sort to models.ProviderSort and it now builds

• #260 c38eade Thanks @AviVahl! - adjust sort type to latest @openrouter/sdk

• #267 89ae694 Thanks @subtleGradient! - Adds support for OpenRouter's BYOK usage accounting. If you have a provider's own API key in your OpenRouter account, cost details are now accessible via usage.costDetails.upstreamInferenceCost -- Thanks @abromberg!