Pull Request Explorer

Summary

Refactor Google interactions option locals

Improves clarity of interaction options for Google language model integration, enhancing maintainability.

Health Assessment

• • Quick refactor with minimal changes, merged within an hour, indicating smooth review and low friction.

Summary

Refactor OpenAI provider options abbreviation

Improves code readability and reduces potential errors in provider option handling.

Health Assessment

• • Minimal changes, quick merge, low risk.

Summary

Refactor Google stream event handling

Improves clarity of local stream event logic, reducing maintenance overhead.

Health Assessment

• • Quick refactor with minimal changes, merged within 48 minutes.

Summary

Refactor Google image delta handling

Improves clarity of image delta logic for Google interactions, enhancing maintainability.

Health Assessment

• • Quick refactor with minimal changes, merged within an hour, indicating smooth review.

Summary

Refactor workflow passthrough stream clarification

Simplifies workflow logic, reducing future maintenance effort.

Health Assessment

• • Quick refactor with minimal changes, merged within an hour.

Summary

Improve build cache naming for clarity

Makes cache key output clearer for developers, aiding debugging and maintenance.

Health Assessment

• • Long cycle time due to delayed review, but minimal code changes and single review round indicate low complexity.

Summary

Extend deployment log retry window

This change extends the log polling window from ~10s to ~15s, reducing failures in CLI integration probes and improving reliability for deployments.

Health Assessment

• • Quick review and merge with minimal changes indicates low complexity and high confidence in the fix.

Summary

Health Assessment

Summary

Upgrade React to latest experimental version

Updates React to the newest experimental release, keeping Next.js compatible with React improvements and bug fixes.

Health Assessment

• • Fast merge with minimal review indicates automated release process.

Summary

Increase test concurrency to 12

Boosts CI throughput by running more tests in parallel, cutting build times while maintaining reliability.

Health Assessment

• • Fast review and merge with minimal changes, indicating low complexity and risk.

Summary

Backport: support document files in Bedrock tool results

Adds support for document files in Amazon Bedrock tool results, ensuring compatibility with older releases.

Health Assessment

• • Fast cycle time and minimal rework indicate a smooth, low-risk backport process.

Summary

Clarify private-cache scope rule

Documentation clarifies private-cache scope guidance and verifies recipe failure modes, improving developer understanding and preventing runtime issues.

Health Assessment

• • The PR had a long review time (~45h) but only a single review round and minimal changes, indicating low technical risk and a straightforward documentation update.

Summary

Version Packages

Automated release of updated package versions to npm, enabling new features and bug fixes for Google and Google Vertex SDKs.

Health Assessment

• • Quick automated release with minimal changes, no review friction.

Summary

Version Packages (canary)

Automated release of canary versions for AI SDK packages, updating changelogs and package.json for Amazon Bedrock, Google, and Google Vertex.

Health Assessment

• • Fast automated release with minimal changes, no rework, quick merge.

Summary

Support document files in Amazon Bedrock tool results

Enables Amazon Bedrock to handle PDF and other document files returned by tools, improving compatibility and user experience.

Health Assessment

• • The PR took almost six days to merge, with a single review after a brief test failure, indicating moderate review friction but a straightforward bug fix.

Summary

Add containerUpload option to pass files to code execution tool

Enables file uploads for Anthropic's code execution tool, supporting CSV, Excel, and JSON files.

Health Assessment

• • The PR has a large scope with 16 files changed, but the review process was relatively smooth with a single review round.

Summary

Remove experimental telemetry warning

Updates documentation to reflect stable telemetry in v7, removing the experimental warning and adjusting code highlighting.

Health Assessment

• • Quick documentation update with minimal review, indicating straightforward change.

Summary

Health Assessment

Summary

Add Speed Insights to vercel-cli skill keywords

Enhances the vercel-cli skill to include Speed Insights and Core Web Vitals, enabling agents to route users to metrics documentation.

Health Assessment

• • Quick review and minimal changes indicate a low‑risk documentation update; the long cycle time reflects a delayed merge rather than technical complexity.

Summary

Version Packages

Automated release of updated package versions, publishing to npm.

Health Assessment

• • Fast cycle time of 14 hours, single commit, minimal review, automated release pipeline ensures quick and low-risk deployment.

Summary

Publish Canary Versions of SDK Packages

Automates publishing of prerelease AI SDK packages to npm, enabling early access and testing.

Health Assessment

• • Quick merge with minimal changes, automated release workflow, low review friction.

Summary

Fix Gemini 3 tool-call replay signature injection

Adds automatic injection of skip_thought_signature_validator for Gemini 3 replays, preventing HTTP 400 errors and improving reliability.

Health Assessment

• • PR took over 4 days to merge with a single commit, indicating a straightforward fix but a slow review process.

Summary

Fix auto-inject skip_thought_signature_validator for Gemini 3

Prevents 400 errors when replaying tool calls with Gemini 3, improving reliability for clients using the Google provider.

Health Assessment

• • The PR added extensive tests and a new validator, but the long cycle time indicates significant review friction and potential impact on delivery timelines.

Summary

Tweak error overlay nav overflow

Improves mobile behavior of error overlay navigation to prevent version badge wrapping.

Health Assessment

• • Quick review and merge process, indicating a straightforward fix.

Summary

Fix time to first output calculation

Accurately measures time to first output, improving performance monitoring and user experience. This ensures reliable metrics for AI responses.

Health Assessment

• • Fast turnaround with minimal review, indicating a straightforward bug fix.

Summary

Backport: fix AIMessageChunk type recognition

Ensures LangChain RemoteGraph streams correctly identify Python AIMessageChunk types, improving compatibility and stream handling.

Health Assessment

• • Quick backport with minimal changes and immediate approval, indicating low complexity and high confidence.

Summary

Upgrade ast-grep

Fixes async closure analysis in ast-grep, improving code quality checks.

Health Assessment

• • Quick fix with minimal changes, resolved in under 5 hours.

Summary

Pass all stream text parts to onChunk

Ensures all streamed text chunks are delivered to the callback, improving consistency for downstream consumers.

Health Assessment

• • Quick fix with minimal review, indicating straightforward bug resolution.

Summary

Fix Vue useChat args forwarding

Ensures useChat methods correctly forward arguments, preventing errors when calling these methods.

Health Assessment

• • Long review cycle indicates low priority or complexity; minimal code changes suggest low risk.

Summary

Fix streamText tests

This PR corrects failing streamText unit tests, ensuring reliable text generation functionality.

Health Assessment

• • Quick merge with minimal changes indicates straightforward test fix.

Summary

Fix incorrect search params caching

Corrects stale search parameters in client-side routing, preventing incorrect URLs and improving navigation consistency for users.

Health Assessment

• • Fast turnaround with a single commit and minimal review, indicating a straightforward bug fix with low complexity and risk.

Summary

Add timeBetweenOutputTokensMs stats

Adds detailed token timing metrics to improve monitoring of AI streaming performance, enabling better UX insights and troubleshooting.

Health Assessment

• • Long cycle time and no review indicate low engagement; large scope suggests high risk of defects and delayed delivery.

Summary

Add edge-light export for OIDC

Adds a new OIDC export for edge-light runtime, enabling token retrieval without Node.js modules, improving authentication support for edge functions.

Health Assessment

• • Fast review and merge with minimal rework, straightforward addition of new export and refactoring of helper modules.

Summary

Rename onFinish to onEnd for SDK

Standardizes callback naming and aggregates final event data, improving consistency and clarity for developers using the AI SDK.

Health Assessment

• • The PR had a long cycle time and many review rounds, indicating significant review friction.
• • Large scope of changes across 20 files and 2400+ lines suggests high maintenance effort.

Summary

Disable package-manager-cache for actions/setup-node

Ensures CI pipelines use pnpm cache correctly, avoiding wasted cache operations and potential build failures.

Health Assessment

• • The PR took over 7 days to merge, with a single review after 24 hours, indicating a slow review process but minimal code changes.

Summary

ServerActionsGraph: pass ModuleGraphLayer as OperationVc

Ensures incremental builds execute the module graph before map_server_actions, preventing unnecessary rebuilds and improving build stability.

Health Assessment

• • Fast review and merge indicate low complexity and high confidence in the change.

Summary

Version Packages (canary)

Automated release of canary versions for multiple SDK packages, updating changelogs and dependencies.

Health Assessment

• • Fast merge with minimal review, typical automated release PR.

Summary

Expose HTTP context on MCPClientError

Adds structured HTTP fields to error for easier downstream handling.

Health Assessment

• • Fast merge with minimal changes and quick review indicates low complexity.

Summary

Wrap model call in telemetry context

Adds telemetry context wrapping for language model calls to ensure proper span nesting, improving observability and preventing regressions in trace data.

Health Assessment

• • The PR fixed a telemetry regression that required multiple commits and a prolonged review cycle, indicating significant rework and potential risk to observability.

Summary

Expose HTTP context on MCPClientError

Adds structured HTTP context to MCPClientError for better error handling and fallback logic.

Health Assessment

• • Quick turnaround with a single commit and fast review; minimal rework and no major blockers.

Summary

Fix CLI release binary workflow

Corrects the release engine to only release binaries via CLI, improving release reliability.

Health Assessment

• • Quick fix with minimal changes, resolved within a day, indicating efficient review and low complexity.

Summary

Update CLI domains section

Adds instructions for using check, price, and buy commands in the Vercel CLI, improving user guidance.

Health Assessment

• • The PR was reviewed and merged within 30 minutes, indicating a straightforward documentation update with minimal impact.

Summary

Resize CI jobs to smaller machines

Reduces CI run time and cost by using smaller runners for tiny jobs.

Health Assessment

• • Long review and merge cycle due to large repo; minimal code changes; low risk.

Summary

Backport: fix Vertex example location

Moves Vertex example to correct folder, ensuring proper organization.

Health Assessment

• • Quick backport with minimal changes, no review needed.

Summary

Version Packages (canary)

Automated release of canary versions for multiple AI SDK packages, updating changelogs and package metadata to trigger npm publishing.

Health Assessment

• • Release PR merged quickly with minimal changes, indicating efficient release process.

Summary

chore: fix Vertex example location

Corrects the file location of a Vertex AI example, ensuring consistency with repository structure and improving discoverability.

Health Assessment

• • Quick fix with minimal changes, no rework, merged within 5 hours.

Summary

Disable CLI auto-update checks for native binary

Prevent auto-update prompts for Vercel native CLI, ensuring native users stay on the binary version.

Health Assessment

• • Quick approval after ~24h review, minimal changes, low risk.

Summary

Speed up deployment scope lookup

This change reduces API calls and latency during deployment, improving user experience and reducing resource usage.

Health Assessment

• • The PR took almost a day to review but required no rework after approval, indicating a straightforward performance improvement.

Summary

Refine error overlay insight pagination

Improves the pagination experience in the dev overlay's Errors/Insights tab, making it more coherent and reliable.

Health Assessment

• • The PR was reviewed and merged quickly, with a cycle time of 5.6 hours, indicating a smooth development process.

Summary

Add AGENTS.md for GitHub Actions

Adds documentation for supported GitHub Actions agents, improving clarity for developers.

Health Assessment

• • The PR received a quick review and required only a single commit, indicating low review friction.
• • Despite the short review time, the overall cycle time was long, suggesting a delayed merge decision.