docs: update v5 announcement page #6526

nicoalbanese · 2025-05-28T13:55:52Z

Summary

Updates:

link to new provider spec
replace continueUntil with stopWhen
add community provider section

Tasks

Formatting issues have been fixed (run pnpm prettier-fix in the project root)

… messages (#5523)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

…nge (#5531) Co-authored-by: Carl Brugger <[email protected]>

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Co-authored-by: Bram Meerten <[email protected]>

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

…rovider options (#5491)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Co-authored-by: Sam Denty <[email protected]> Co-authored-by: Sam Denty <[email protected]>

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

## Background The below MCP SSE message will make the stream reader timeout. ``` data: /message?sessionId=123\r\nevent: endpoint ``` ## Summary Fix the `splitLines` parser issue. The CRLF pointer was incorrectly implemented.

# Releases ## [email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - d87b9d1: fix(provider-utils): fix SSE parser bug (CRLF) ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - [email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [d87b9d1] - @ai-sdk/[email protected] - @ai-sdk/[email protected] Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

…ming stream (#6224) ## Background This pull request updates the documentation for the GET handler in the resumable streams section to account for a race condition that may occur when the stream ends in-flight. ## Summary Updated the route by adding the code snippet that handles the above condition. ## Tasks - [x] Documentation has been added / updated (for bug fixes / features) - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root) ## Related Issues #6052

## Background Bug with Vue that led to status not updating when tab was changed. ## Summary Changes status from using SWR to using Vue ref.

# Releases ## @ai-sdk/[email protected] ### Patch Changes - b54cb59: fix (ai-sdk/vue): fix status reactivity Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

## Background In mcp doc, `mcpClient.tools()` need to await ## Verification Just fix doc problem Co-authored-by: georgezouq <[email protected]> Co-authored-by: Lars Grammel <[email protected]>

## Background [Vertex now supports extraction of thinking tokens in certain Gemini models](https://cloud.google.com/vertex-ai/generative-ai/docs/thinking). When the configuration is passed via `providerOptions`, the sdk: 1. Did not extract reasoning tokens 2. Did not pass `include_thoughts` to the provider ## Summary Added extraction logic to google-generative-ai package to parse reasoning tokens. Added a `includeThoughts` switch to the `thinkingConfig` for vertex models. ## Verification I verified it manually. Testable via examples/ai-core/src/stream-text/google-vertex-reasoning.ts. Easily copiable to google provider. ## Related Issues Fixes #6259

## Background Release requires change to `google` provider package. ## Summary Add changeset.

# Releases ## @ai-sdk/[email protected] ### Patch Changes - 4b2e1b0: Add reasoning token output support for gemini models via Vertex AI Provider ## @ai-sdk/[email protected] ### Patch Changes - fe24216: Add reasoning token output support for gemini models via Vertex AI Provider - Updated dependencies [4b2e1b0] - @ai-sdk/[email protected] Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

## Background Like many other OpenTelemetry instrumentations, Laminar depends on `@opentelemetry/instrumentation`, which, in turn, depends on `require-in-the-middle` and `import-in-the-middle`. Importing and initializing Laminar inside Next.js `instrumentation.ts` file causes Next.js to try resolving these two packages, but fails, and results in: - Laminar not being able to send traces. This is because unlike many other instrumentation libraries, Laminar is not intrusive and does not set its tracer provider globally (so that others, e.g. `@vercel/otel` can set theirs). - Error messages (see below) We have tried many different things to debug, including bundling Laminar differently, shipping those two packages within Laminar as `noExternal`, adding a separate entrypoint in our package for Next.js, but nothing seems to have worked. The only thing that's worked was adding `@lmnr-ai/lmnr` in `serverExternalPackages` in [next.config](https://nextjs.org/docs/app/api-reference/config/next-config-js/serverExternalPackages). ## Summary Add a subsection within the Next.js section that describes

## Background It is not clear what AI SDK version user-submitted issues refer to. ## Summary Add version section to issue template.

## Background Integrating with Patronus! ## Summary A new Patronus.mdx file showing how logs can be imported into Patronus via OTel. ## Tasks - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root) --------- Co-authored-by: Snigdha Banda <[email protected]> Co-authored-by: nicoalbanese <[email protected]> Co-authored-by: Nico Albanese <[email protected]>

## Summary Announce v5 Alpha ## Tasks - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root)

…spans (#6357) ## Background `generateObject`, `generateText`, `streamText`, and `streamObject` currently call `JSON.stringify` on the input messages. If the input messages contain an image, it is most likely normalized into a `Uint8Array`. `JSON.stringify` does not the most obvious things to TypedArrays including `Uint8Array`. ```javascript // this returns '{"0": 1,"1": 2,"2": 3}', where I'd expect this to be '[1,2,3]' JSON.stringify(new Uint8array([1, 2, 3])) ``` In practice, this results in bloating images by about 5-15x depending on the original image size. For Laminar, for example, a span with 3 avg sized images will not be able to be sent as it is larger than the (reasonably high) gRPC payload size for our traces endpoint. From [MDN docs](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/JSON/stringify#examples): ```javascript // TypedArray JSON.stringify([new Int8Array([1]), new Int16Array([1]), new Int32Array([1])]); // '[{"0":1},{"0":1},{"0":1}]' JSON.stringify([ new Uint8Array([1]), new Uint8ClampedArray([1]), new Uint16Array([1]), new Uint32Array([1]), ]); // '[{"0":1},{"0":1},{"0":1},{"0":1}]' JSON.stringify([new Float32Array([1]), new Float64Array([1])]); // '[{"0":1},{"0":1}]' ``` ## Summary Added a function that maps over messages in a `LanguageModelV1Prompt` and maps over content parts in each message, replacing `UInt8Array`s with raw base64 strings instead. Call this function when calling `recordSpan` for the inner (doStream/doGenerate) span in `generateObject`, `generateText`, `streamText`, and `streamObject`. ## Verification Ran this small script against a local instance of Laminar and logged the Telemetry payloads (span attributes) on the backend to verify that they are indeed base64. ```javascript import { Laminar, getTracer } from '@lmnr-ai/lmnr' Laminar.initialize(); import { openai } from '@ai-sdk/openai' import { generateText, generateObject, streamText, streamObject, tool } from "ai"; import { z } from "zod"; import dotenv from "dotenv"; dotenv.config(); const handle = async () => { const imageUrl = "https://upload.wikimedia.org/wikipedia/commons/b/bc/CoinEx.png" const imageData = await fetch(imageUrl) .then(response => response.arrayBuffer()) .then(buffer => Buffer.from(buffer).toString('base64')); const o = streamObject({ schema: z.object({ text: z.string(), companyName: z.string().optional().nullable(), }), messages: [ { role: "user", content: [ { type: "text", text: "Describe this image briefly" }, { type: "image", image: imageData, mimeType: "image/png" } ] } ], model: openai("gpt-4.1-nano"), experimental_telemetry: { isEnabled: true, tracer: getTracer() } }); for await (const chunk of o.fullStream) { console.log(chunk); } await Laminar.shutdown(); }; handle().then((r) => { console.log(r); }); ``` ## Related Issues Fixes #6210

# Releases ## [email protected] ### Patch Changes - ed0ebeb: Avoid JSON.strinfigy on UInt8Arrays for telemetry ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [ed0ebeb] - [email protected] Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

## Background The Codex VM uses pnpm 10. I am running into build issues with Codex and want to rule out that they are caused by pnpm version differences. ## Summary Upgrade pnpm to version 10.

## Background Vercel is adding an API for the v0 model per https://vercel.com/docs/v0/api and an AI SDK provider would make it easier for developers to interact with it. ## Summary Added an initial provider for the Vercel API. ## Verification Added examples across several features and executed end to end manually. Co-authored-by: Walter Korman <[email protected]>

This PR was opened by the [Changesets release](https://github.com/changesets/action) GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm 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 ## @ai-sdk/[email protected] ### Patch Changes - 42e37fb: feat (provider/vercel): initial vercel provider Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

## Background This pull request adds support for Anthropic's new Claude v4 models. ## Summary Updated the model ids to include `claude-4-opus-20250514` and `claude-4-sonnet-20250514`. ## Tasks - [x] A _patch_ changeset for relevant packages has been added (for bug fixes / features - run `pnpm changeset` in the project root) - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root)

# Releases ## @ai-sdk/[email protected] ### Patch Changes - f64f4f0: feat (providers/anthropic): add claude v4 models ## @ai-sdk/[email protected] ### Patch Changes - Updated dependencies [f64f4f0] - @ai-sdk/[email protected] Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

## Summary Adds guide for Claude 4 models. ## Tasks - [ ] Tests have been added / updated (for bug fixes / features) - [ ] Documentation has been added / updated (for bug fixes / features) - [ ] A _patch_ changeset for relevant packages has been added (for bug fixes / features - run `pnpm changeset` in the project root) - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root)

shaper · 2025-05-30T16:09:34Z

content/docs/01-announcing-ai-sdk-5-alpha/index.mdx

@@ -76,6 +88,8 @@ LanguageModelV2 represents a complete redesign of how the AI SDK communicates wi
 - **Improved Type Safety** - The new LanguageModelV2 provides better TypeScript type guarantees, making it easier to work with different content types
 - **Simplified Extensibility** - Adding support for new model capabilities no longer requires changes to the core structure

+To learn more about LanguageModelV2, visit the [LanguageModelV2 documentation](https://v5.ai-sdk.dev/providers/community-providers/custom-providers).


wrap LMv2 in backticks

## Summary Adds v5 alpha announcement page to v5 branch. ## Tasks - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root) ## Future Work add updates from #6526

dylanmoz and others added 30 commits April 3, 2025 08:10

docs: fix typo (#5521)

12ac4c3

docs: fix highlights in node.js guide (#5519)

fba2bd5

docs: update baseten community provider docs (#5517)

1a02790

fix (provider/google-vertex): fix anthropic support for image urls in…

292f543

… messages (#5523)

Version Packages (#5525)

8362d21

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

fix (core): send buffered text in smooth stream when stream parts cha…

c45d100

…nge (#5531) Co-authored-by: Carl Brugger <[email protected]>

Version Packages (#5532)

4e28bba

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

chore (ui/react,vue): update more tests to unified test server (#5509)

2c19b9a

docs: Use new create-nuxt tool when starting Nuxt projects (#5175)

9cfaaf0

feat (docs): Helicone observability (#5472)

bc46777

docs: add local caching middleware recipe (#5540)

0ee98df

fix (docs): highlight (#5545)

5952cfb

feat (core): Add finishReason field to NoObjectGeneratedError (#5541)

772a2d7

Co-authored-by: Bram Meerten <[email protected]>

Version Packages (#5536)

ed79357

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

fix (docs): troubleshooting page description (#5551)

2abf7a6

feat (provider/{google, google-vertex}): expose type for validating p…

1789884

…rovider options (#5491)

fix (docs): correct content outside description in jest docs (#5559)

5f61296

Version Packages (#5558)

bd8b669

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

feat (core): add chunking functions support to smoothStream (#5548)

3d1bd38

Co-authored-by: Sam Denty <[email protected]> Co-authored-by: Sam Denty <[email protected]>

Version Packages (#5561)

5c47bbd

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

fix (core): consume stream on abort (#5492)

665a567

feat (provider/groq): add llama 4 model (#5565)

8fdc8fe

Version Packages (#5564)

64f37ac

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

fix (provider/google): allow "OFF" for Google HarmBlockThreshold (#5550)

1e8e66d

feat (provider/azure): add OpenAI responses API support (#5461)

e82024e

Version Packages (#5570)

4b97ea9

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

feat (providers/deepinfra): add llama 4 models (#5572)

264b1e0

Version Packages (#5573)

e5f6a50

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

fix (ai/mcp): prevent mutation of customEnv (#5583)

3e88f4d

docs: add stackone toolset (#5585)

b8e7fb7

twxia and others added 23 commits May 7, 2025 13:58

fix(provider-utils): fix SSE parser bug (CRLF) (#6197)

d87b9d1

## Background The below MCP SSE message will make the stream reader timeout. ``` data: /message?sessionId=123\r\nevent: endpoint ``` ## Summary Fix the `splitLines` parser issue. The CRLF pointer was incorrectly implemented.

fix (ai-sdk/vue): status reactivity (#6234)

b54cb59

## Background Bug with Vue that led to status not updating when tab was changed. ## Summary Changes status from using SWR to using Vue ref.

Version Packages (#6249)

396d013

# Releases ## @ai-sdk/[email protected] ### Patch Changes - b54cb59: fix (ai-sdk/vue): fix status reactivity Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

fix (docs): mcpClient.tools() need to await (#6186)

d79f373

## Background In mcp doc, `mcpClient.tools()` need to await ## Verification Just fix doc problem Co-authored-by: georgezouq <[email protected]> Co-authored-by: Lars Grammel <[email protected]>

feat(providers/google): Add reasoning token output support (#6284)

4b2e1b0

## Background Release requires change to `google` provider package. ## Summary Add changeset.

chore (github): add version section to support issue template (#6315)

c7aa165

## Background It is not clear what AI SDK version user-submitted issues refer to. ## Summary Add version section to issue template.

docs: add v5 alpha announcement (#6344)

c97ae1a

## Summary Announce v5 Alpha ## Tasks - [x] Formatting issues have been fixed (run `pnpm prettier-fix` in the project root)

fix (docs): update alpha (#6356)

bdd06b5

chore: upgrade to pnpm 10 (#6395)

48ef311

## Background The Codex VM uses pnpm 10. I am running into build issues with Codex and want to rule out that they are caused by pnpm version differences. ## Summary Upgrade pnpm to version 10.

docs: update v5 announcement page

15c1dea

github-actions bot assigned nicoalbanese May 28, 2025

vercel bot deployed to Preview May 28, 2025 13:56 View deployment

nicoalbanese mentioned this pull request May 28, 2025

docs (v5): Add v5 alpha announcement page #6527

Merged

1 task

shaper approved these changes May 30, 2025

View reviewed changes

lgrammel approved these changes Jun 1, 2025

View reviewed changes

lgrammel force-pushed the main branch from e75d6fc to 9aa4214 Compare July 2, 2025 08:41

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

docs: update v5 announcement page #6526

docs: update v5 announcement page #6526

Uh oh!

nicoalbanese commented May 28, 2025

Uh oh!

shaper May 30, 2025

Uh oh!

Uh oh!

docs: update v5 announcement page #6526

Are you sure you want to change the base?

docs: update v5 announcement page #6526

Uh oh!

Conversation

nicoalbanese commented May 28, 2025

Summary

Tasks

Uh oh!

shaper May 30, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!