43 Releases published by 1 person

• ai@5.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/elevenlabs@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/assemblyai@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/deepgram@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/deepseek@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/codemod@2.0.0-beta.0

  published Jun 24, 2025

• @ai-sdk/cohere@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/anthropic@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/amazon-bedrock@3.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/cerebras@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/azure@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/luma@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/google-vertex@3.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/hume@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/gladia@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/fal@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/deepinfra@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/groq@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/gateway@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/fireworks@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/provider@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/openai-compatible@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/xai@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/replicate@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/llamaindex@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/openai@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/provider-utils@3.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/perplexity@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/lmnt@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/google@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/mistral@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/langchain@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/react@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/revai@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/togetherai@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/svelte@3.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/valibot@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/vercel@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/rsc@1.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/vue@2.0.0-beta.1

  published Jun 24, 2025

• @ai-sdk/google-vertex@3.0.0-beta.2

  published Jun 25, 2025

• @ai-sdk/google@2.0.0-beta.2

  published Jun 25, 2025

• @ai-sdk/anthropic@2.0.0-beta.2

  published Jun 25, 2025

37 Pull requests merged by 5 people

• chore (providers/gateway): update language model ids

  #6883 merged Jun 26, 2025

• fix (provider/google): remove non-functional models

  #6874 merged Jun 26, 2025

• fix (provider-utils): detect failed fetch in browser environments

  #6871 merged Jun 26, 2025

• chore (provider-utils): update eventsource-parser to 3.0.3

  #6869 merged Jun 26, 2025

• Version Packages (beta)

  #6858 merged Jun 25, 2025

• fix (provider/anthropic): send tool call id in tool-input-start chunk

  #6859 merged Jun 25, 2025

• fix (provider/google): omit system message for gemma models

  #6857 merged Jun 25, 2025

• chore: update v5 beta announcement page on main

  #6850 merged Jun 24, 2025

• docs (v5): add migration guide

  #6791 merged Jun 24, 2025

• docs (v5): update alpha page to beta

  #6837 merged Jun 24, 2025

• Version Packages (beta)

  #6848 merged Jun 24, 2025

• chore (release): switch to beta

  #6847 merged Jun 24, 2025

• feat (ai): expose http chat transport type

  #6845 merged Jun 24, 2025

• feat: add transcription and speech model support to provider registry

  #6843 merged Jun 24, 2025

• spec (ai): add provider options to tools

  #6844 merged Jun 24, 2025

• fix: anthropic computer tool

  #6841 merged Jun 24, 2025

• feat (ui): add state for text and reasoning ui message parts

  #6842 merged Jun 24, 2025

• feat (provider): support changing provider, model, supportedUrls in middleware

  #6840 merged Jun 24, 2025

• feat (provider): support arbitrary media types in tool results

  #6839 merged Jun 24, 2025

• feat (provider): add name for provider defined tools for future validation

  #6838 merged Jun 24, 2025

• chore (provider/openai): switch default to openai responses api

  #6833 merged Jun 23, 2025

• feat: provider-executed tools

  #6815 merged Jun 23, 2025

• refactoring: refactor the server side tool definitions

  #6827 merged Jun 23, 2025

• chore: add language setting to speechv2

  #6826 merged Jun 23, 2025

• fix (ai): add tests and examples for openai responses

  #6825 merged Jun 23, 2025

• chore (providers/gateway): update language model ids

  #6828 merged Jun 23, 2025

• refactoring (provider): collapse provider defined tools into single definition

  #6821 merged Jun 22, 2025

• refactoring: move tools helper into provider-utils

  #6820 merged Jun 22, 2025

• feat (ai): add output schema for tools

  #6819 merged Jun 22, 2025

• refactoring: introduce FlexibleSchema

  #6818 merged Jun 22, 2025

• feat (ai): add InferToolInput and InferToolOutput helpers

  #6817 merged Jun 22, 2025

• chore: add tests and examples for streaming

  #6816 merged Jun 22, 2025

• feat: forward id, streaming start, streaming end of content blocks

  #6812 merged Jun 22, 2025

• feat(openai): add file search tool support

  #6802 merged Jun 20, 2025

• feat (provider/anthropic): send web search tool calls

  #6807 merged Jun 20, 2025

• feat (provider/anthropic): parse websearch tool args

  #6809 merged Jun 20, 2025

• refactoring (ai): restructure provider-defined tools

  #6808 merged Jun 20, 2025

12 Pull requests opened by 12 people

• fix: use appropriate domain for vertex global

  #6810 opened Jun 20, 2025

• feat(docs): add Claude Code community provider

  #6813 opened Jun 20, 2025

• fix(amazon-bedrock): update amazon.nova-canvas-v1:0 documentation

  #6836 opened Jun 24, 2025

• fix(ai): selectTelemetryAttributes attribute's value should ignore `u…

  #6846 opened Jun 24, 2025

• feat(openai-compatible): passing config to chat models

  #6849 opened Jun 24, 2025

• feat (providers/gateway): improve oidc api key client auth flow

  #6851 opened Jun 24, 2025

• added JigsawStack as a listed tool provider for AI SDK

  #6853 opened Jun 24, 2025

• Add Vercel SDK Provider list

  #6863 opened Jun 25, 2025

• Version Packages (beta)

  #6870 opened Jun 26, 2025

• feat(vue): add experimental_resume feature

  #6880 opened Jun 26, 2025

• docs (v5): Update AI SDK migration guide with Zod dependency

  #6882 opened Jun 26, 2025

• feature: using Zod 4 for internal stuff

  #6884 opened Jun 26, 2025

191 Issues closed by 15 people

• Feature request: `generateObject` support for AI Gateway

  #6877 closed Jun 26, 2025

• LangChainAdapter DataStreamProtocol changes not parsing to Messages properly

  #5650 closed Jun 26, 2025

• tool execute inputs no longer typed with zod 3.24

  #6876 closed Jun 26, 2025

• streamText terminates prematurely, stopping generation

  #6875 closed Jun 26, 2025

• Provide Options to Remove Duplicated Data in Message Object

  #5342 closed Jun 26, 2025

• Vertex - Use grounding with my own data.

  #4679 closed Jun 26, 2025

• Tool Calling with Search Grounding appears to not work `(gemini-2.0-flash)`

  #6219 closed Jun 26, 2025

• Vertex own models do not work out of the box

  #5605 closed Jun 26, 2025

• StreamText Broken Using @ai-sdk/google-vertex in Version 2 and above

  #6162 closed Jun 26, 2025

• Tool calling not working with Gemini when useSearchGrounding is set to true

  #6302 closed Jun 26, 2025

• Temperature Not Supported by Model `codex-mini-latest` with `generateText` and `temperature` set to `undefined`

  #6789 closed Jun 26, 2025

• OpenAI Reasoning Summary parts aren't split

  #6742 closed Jun 26, 2025

• Claude 3.7 may require specific newline handling after tool-calls

  #6630 closed Jun 26, 2025

• Is using `experimental_attachments` required? `file` `parts` aren't working.

  #6623 closed Jun 26, 2025

• add AI/ML API as provider

  #6530 closed Jun 26, 2025

• providerMetadata is empty

  #6489 closed Jun 26, 2025

• Runware AI provider support

  #6478 closed Jun 26, 2025

• v5 mutates message parts which causes problems with the React Compiler

  #6466 closed Jun 26, 2025

• Message Parts contain only one "tool-invocation" type, the latest instead of all the tool invocations in order.

  #6342 closed Jun 26, 2025

• New provider for Meta

  #6242 closed Jun 26, 2025

• [v5 and v4 ] StreamText does not have raw usage

  #6702 closed Jun 26, 2025

• `experimental_continueSteps` causes the last chunk to arrive after the tool call

  #6125 closed Jun 26, 2025

• `experimental_prepareRequestBody` is not compatible with regular `body`

  #6734 closed Jun 26, 2025

• Feature Request: Expose Unified AI Providers

  #6056 closed Jun 26, 2025

• support image to image generation?

  #6046 closed Jun 26, 2025

• Exception in tool call fail with a generic An error occurred

  #6039 closed Jun 26, 2025

• `streamText` -> `textStream` doesn't offer an ability to surface an error before reading

  #6018 closed Jun 26, 2025

• feature - Support build-in browser models apis

  #5953 closed Jun 26, 2025

• useChat sending message to Google ADK API

  #5878 closed Jun 26, 2025

• import_provider2.APICallError.isAPICallError is not a function

  #5781 closed Jun 26, 2025

• Improve performance of eventsource parser

  #5862 closed Jun 26, 2025

• Support continue generation when tool calling fails

  #5747 closed Jun 26, 2025

• Can we make onStepFinish do more?

  #5743 closed Jun 26, 2025

• Remove tools schema from request when reaching last step

  #5679 closed Jun 26, 2025

• support to stop further completion after tool call

  #5655 closed Jun 26, 2025

• Field content is empty string when a tool is invoked.

  #5616 closed Jun 26, 2025

• temperature = 0 is always passed to gpt-4o-search-preview which breaks it

  #5609 closed Jun 26, 2025

• AI Initialization Failure, `streamText` Not Generating Output, and Long Request Time

  #5597 closed Jun 26, 2025

• LangChainAdapter uses deprecated StreamData() method

  #5552 closed Jun 26, 2025

• providerOptions on embed, embedMany

  #5664 closed Jun 26, 2025

• HarmonyLinkingError: export 'useId' (imported as 'useId2') was not found in 'react'

  #5515 closed Jun 26, 2025

• Dynamic Control of experimental_activeTools

  #5478 closed Jun 26, 2025

• Improve documentation: Expected response structure for useChat custom endpoints

  #5462 closed Jun 26, 2025

• Error: Maximum update depth exceeded. Only when 'initialMessages' props is passed in 'useChat' react hook.

  #5428 closed Jun 26, 2025

• @ai-sdk/solid: New chat message forces the chat window to scroll to the top

  #5411 closed Jun 26, 2025

• Error: Attempted import error: 'parseProviderOptions' is not exported from '@ai-sdk/provider-utils' (imported as 'parseProviderOptions').

  #5369 closed Jun 26, 2025

• Adding a tool to streamText breaks messages using Ollama provider / Assistant-UI

  #5330 closed Jun 26, 2025

• Get raw finish reason from provider

  #5307 closed Jun 26, 2025

• Returning texts and images from latest gemini-2.0-flash-exp-image-generation

  #5302 closed Jun 26, 2025

• Order of message parts get mixed in onFinish()

  #5275 closed Jun 26, 2025

• Tool is Invoked Multiple Times Based on maxSteps Value

  #5195 closed Jun 26, 2025

• Long tool call IDs break OpenAI's validation

  #5171 closed Jun 26, 2025

• streamObject using fastify backend does not seem to work

  #5160 closed Jun 26, 2025

• Support useChat Without an API Route (Client-side usage)

  #5140 closed Jun 26, 2025

• @ai-sdk/perplexity sources support

  #5123 closed Jun 26, 2025

• Add SambaNova System's cloud API as a provider and build a new package for it

  #5112 closed Jun 26, 2025

• Data stream improvement: message specific data stream

  #5046 closed Jun 26, 2025

• React useChat() doesn't stop stream and reset status when `id` is changed

  #5015 closed Jun 26, 2025

• Client-Side Support for useChat with Ollama API Support and Ability to Control baseUrl on Client-Side

  #4984 closed Jun 26, 2025

• Header: 'anthropic-beta': 'interleaved-thinking-2025-05-14' Causes error "...`tool_use` ids were found without `tool_result` blocks immediately after: ..."

  #6472 closed Jun 26, 2025

• Hookable Control Over Agent Loop Iterations in AI SDK

  #4954 closed Jun 26, 2025

• Option to split validation schema and prompt schema

  #4924 closed Jun 26, 2025

• Remove non-functional Google AI model IDs from GoogleGenerativeAIModelId type

  #6481 closed Jun 26, 2025

• (vue) useChat not responsive to changes in the conversation ID

  #4900 closed Jun 26, 2025

• useChat not working with SWR 2.3.2

  #4836 closed Jun 26, 2025

• SiliconFlow API provider support

  #4785 closed Jun 26, 2025

• Transform response before parsing JSON

  #4710 closed Jun 26, 2025

• Property '_internal' does not exist on type...

  #4686 closed Jun 26, 2025

• handleSubmit only submits the first letter

  #4534 closed Jun 26, 2025

• React useChat(): experimental_attachments doesn't work for reload() and append()

  #4519 closed Jun 26, 2025

• External URL Handling Error for Non-Image contentType in experimental_attachments

  #4443 closed Jun 26, 2025

• Standardize on `contentType` or `mimeType` but not both

  #4360 closed Jun 26, 2025

• Chat example with tools in Docs broken

  #4229 closed Jun 26, 2025

• Provide additional information about request in generateText errors

  #4193 closed Jun 26, 2025

• Support Dify Provider

  #4084 closed Jun 26, 2025

• toolChoice: 'required' and { toolName: 'x', type: 'tool' } send ai vercel SDK into endless loop when using streamText

  #3944 closed Jun 26, 2025

• Feature Request: Cost Calculation from usage (Based on Provider + Model prices)

  #3932 closed Jun 26, 2025

• support openai SSEs in useChat

  #3793 closed Jun 26, 2025

• The v3.4 to v4 migration codemod is unusably slow

  #3766 closed Jun 26, 2025

• keepLastMessageOnError does not keep last message on DataStream

  #3579 closed Jun 26, 2025

• Provide a safeparse / way to access the result in case of zod schema validation error

  #3451 closed Jun 26, 2025

• Return the execution result for the provider execution tools

  #3439 closed Jun 26, 2025

• Tool with async fetch doesn't return data

  #3372 closed Jun 26, 2025

• Modify the AI Message & properties being sent to the UI via data stream

  #3357 closed Jun 26, 2025

• experimental_activeTools needs a hallucination filter when tools are disabled mid-conversation

  #3322 closed Jun 26, 2025

• Using the same id with useChat is not hooking into the existing session

  #3266 closed Jun 26, 2025

• `useObject` should support onChunk callback option.

  #5965 closed Jun 26, 2025

• Custom provider name when extending the openai provider

  #3209 closed Jun 26, 2025

• useChat() streaming working in localhost but not docker deployment with nginx

  #3159 closed Jun 26, 2025

• Expose request body for all steps in generateText

  #3157 closed Jun 26, 2025

• Add Google's embedding and caching to vertex ai

  #3128 closed Jun 26, 2025

• Minor bug in Next.js app router docs example code

  #3126 closed Jun 26, 2025

• Why not a single client side library rather than a separate for nuxt/react/svelte etc

  #3121 closed Jun 26, 2025

• Allow vertex to work on Vercel-Edge

  #3096 closed Jun 26, 2025

• Send `data` or custom fields from `streamText` to `useChat` similar to `annotations`

  #3056 closed Jun 26, 2025

• Support dangerouslyAllowBrowser for Anthropic

  #3041 closed Jun 26, 2025

• Support for Additional Inference Providers (DeepInfra, Hyperbolic) and Voyage AI for Embeddings

  #2996 closed Jun 26, 2025

• ai-sdk/ui: Provide a way to customise the user message in useChat

  #2987 closed Jun 26, 2025

• Extend messages type

  #2932 closed Jun 26, 2025

• Solid: useObject support

  #2904 closed Jun 26, 2025

• Sveltekit handleSubmit sends back data as undefined

  #2796 closed Jun 26, 2025

• Add JigsawStack SDK as part of the AI SDK

  #2787 closed Jun 26, 2025

• Add Ai21Lab Provider @ai-sdk/ai21 for Jamba 1.5 Models

  #2783 closed Jun 26, 2025

• Using a schema with useChat

  #2771 closed Jun 26, 2025

• createStructuredFewShot prompt helper

  #2696 closed Jun 26, 2025

• How to Call API Request without using handleSubmit and add More inputs

  #2869 closed Jun 26, 2025

• Solid.js: attachment support for useChat

  #2901 closed Jun 26, 2025

• Add toolCallId and toolName to stream part of type "error"

  #2578 closed Jun 26, 2025

• How can i use moderation api from open ai with vercel ai sdk, there was nothing about this in vercel ai docs

  #2555 closed Jun 26, 2025

• Leftover '@ai-sdk/provider@0.0.11' after installing @ai-sdk/anthropic@latest in certain environments

  #2537 closed Jun 26, 2025

• HuggingFaceStream is giving stream not compatible with StreamingTextResponse

  #2485 closed Jun 26, 2025

• Support `{ type: 'text', url: ... }` chunks

  #2406 closed Jun 26, 2025

• Make `Message` `createdAt` key required

  #2184 closed Jun 26, 2025

• Please export the type for onFinish callback on streamText

  #2065 closed Jun 26, 2025

• Chunk size increases for long streams

  #1990 closed Jun 26, 2025

• Streamed response rendering out of order

  #1892 closed Jun 26, 2025

• Support for Passing Arbitrary Extra Metadata in API Requests

  #1881 closed Jun 26, 2025

• Support for Custom Roles in convertToLanguageModelMessage

  #1771 closed Jun 26, 2025

• env.local env does not override shell env

  #1714 closed Jun 26, 2025

• Erratic behavior of useChat - Could SWR be the issue?

  #1656 closed Jun 26, 2025

• Expose a `AssistantStream` function

  #1432 closed Jun 26, 2025

• handleSubmit not being called conditionally

  #1394 closed Jun 26, 2025

• examples/next-huggingface doesn't work.

  #1338 closed Jun 26, 2025

• Generic Messages

  #1180 closed Jun 26, 2025

• useChat setMessages Assertion

  #1106 closed Jun 26, 2025

• Metadata with assistants

  #999 closed Jun 26, 2025

• Maintain data over multiple sessions.

  #976 closed Jun 26, 2025

• ability to transform response from backend in useChat

  #975 closed Jun 26, 2025

• experimental_StreamData，If the string is too long, it will be truncated.

  #962 closed Jun 26, 2025

• useChat key parameter nonfunctional

  #896 closed Jun 26, 2025

• [feature] provide a callback to useChat for when messages are completed

  #501 closed Jun 26, 2025

• isLoading (useChat) gets stuck on "true" if id changes before streaming starts

  #819 closed Jun 26, 2025

• use multiple useChat methods stop function not work

  #486 closed Jun 26, 2025

• When there is an ongoing AI streaming, switch the chat id, will cause data overwrite.

  #696 closed Jun 26, 2025

• Remove default value `temperature=0` param from the middleware in ai core

  #6034 closed Jun 26, 2025

• Add feature to continue streaming on browser refresh/exit page

  #470 closed Jun 26, 2025

• Support client-only use of chat library

  #208 closed Jun 26, 2025

• useAssistant: option to ignore annotations in text deltas

  #2215 closed Jun 26, 2025

• setThreadId is changing frequently

  #2721 closed Jun 26, 2025

• useAssistant: Get tool returns on client side.

  #3089 closed Jun 26, 2025

• Include the `createdAt` field in `AssistantResponse` and `useAssistant`

  #3129 closed Jun 26, 2025

• useChat to support streamText with structured output

  #4277 closed Jun 26, 2025

• Use tools with useAssistant() OR enable code-interpreter on useChat()

  #3852 closed Jun 26, 2025

• useAssistant hook createMessage method

  #3016 closed Jun 26, 2025

• Tweak documentation for AI SDK UI -> OpenAI Assistants

  #2735 closed Jun 26, 2025

• Svelte useAssistant stop function does not cancel run

  #1743 closed Jun 26, 2025

• `onError` hook not called when an error occurs during streaming in `useAssistant`

  #1658 closed Jun 26, 2025

• [AsistantResponse] - to save streaming data to external database

  #1461 closed Jun 26, 2025

• Annotations array empty with OpenAI and AssistantResponse and useAssistant

  #1431 closed Jun 26, 2025

• new assistant file upload api support

  #1378 closed Jun 26, 2025

• add props onFinish or onCompletion in experimental_useAssistant

  #1295 closed Jun 26, 2025

• Support initial input as an option for useAssistant api

  #1250 closed Jun 26, 2025

• Assistant Hook: Add client side function handling and other events

  #868 closed Jun 26, 2025

• Final run has not been received

  #4361 closed Jun 26, 2025

• Error doesn't clear on useAssistant

  #4615 closed Jun 26, 2025

• "skip" or "enabled" parameter to "useAssistant" hook.

  #5219 closed Jun 26, 2025

• AssistantResponse from nestjs not sending stream to the client

  #6156 closed Jun 26, 2025

• failed fetches not mapped to APICallError

  #6370 closed Jun 26, 2025

• bug: infinite re-render loop because missing `useCallback` in `useChat => handleInputChange`

  #6266 closed Jun 26, 2025

• This URL is disallowed by the website's robots.txt file

  #6279 closed Jun 26, 2025

• define tool response schema

  #6298 closed Jun 26, 2025

• Mistral object generation fails with `mode: tool`

  #6330 closed Jun 26, 2025

• Tool call is broken

  #6333 closed Jun 26, 2025

• Option to Remove think part from Reasoning Models for streamText

  #6362 closed Jun 26, 2025

• When setting multi-step generateText, promptMessages should be compressed or partially truncated

  #6714 closed Jun 26, 2025

• Nested agents required for agent swarm/orchestration doesn't work

  #6381 closed Jun 26, 2025

• Add support for server_tool_use

  #6401 closed Jun 26, 2025

• [v5] Implement `.providerMetadata` across all image models

  #6405 closed Jun 26, 2025

• Methods `writeSource` and `writeMessageAnnotation` do not update data of `sources` and `messages[i].annotations` in `StreamTextOnFinishCallback`

  #6433 closed Jun 26, 2025

• [Bug]: Metro resolution failed in React Native (Expo) for "eventsource-parser/stream"

  #6862 closed Jun 26, 2025

• Does `api` exist on `useChat()` for ai 5.0.0-beta.1?

  #6866 closed Jun 25, 2025

• Support for Orchestrating Agents: Routines and Handoffs like it was done in swarm

  #3233 closed Jun 25, 2025

• `message.id` changes when new `part` is added

  #5318 closed Jun 25, 2025

• Message Concatenation Issue with Multiple Roundtrips

  #4654 closed Jun 25, 2025

• Add support for Anthropic web search tool

  #6666 closed Jun 25, 2025

• Support for Zod v4

  #5682 closed Jun 25, 2025

• Generate assistant completion after tool execution

  #6508 closed Jun 25, 2025

• streamText resulting text is not blocking

  #6522 closed Jun 25, 2025

• `messageAnnotation` from tool `execute` (via `DataStreamWriter`) not streaming to client before tool step completion

  #6539 closed Jun 25, 2025

• support flex processing for OpenAI

  #6545 closed Jun 25, 2025

• Bug: Random file names break prompt cache effectiveness

  #6555 closed Jun 25, 2025

• Feature Request: Add endTool behavior to terminate agent run after specific tool is called

  #6616 closed Jun 25, 2025

• How to change model between multiple steps?

  #6619 closed Jun 25, 2025

• Missing Reasoning Chain in streamText Response

  #6625 closed Jun 25, 2025

• [Feature Request] Add native citation support when using Cohere's Command A model

  #6628 closed Jun 25, 2025

• [@ai-sdk/google-vertex] Unexpected global location domain

  #6644 closed Jun 25, 2025

• [V5] Docs client side crash when searching "chat store"

  #6703 closed Jun 25, 2025

• Bug/Feature request: add a message part for tool-invocation before the creation of the tool's input

  #6707 closed Jun 25, 2025

• defaultChatStore not exported in ai@5.0.0-alpha.13

  #6758 closed Jun 25, 2025

• experimental_resume is calling api/chat?chatId=undefined despite having set useChat(id)

  #6407 closed Jun 24, 2025

• Stream Text Hangs Forever when abort signal is already aborted

  #6814 closed Jun 23, 2025

16 Issues opened by 15 people

• unhandled promise rejection in process-ui-message-stream

  #6879 opened Jun 26, 2025

• AI SDK v5 beta docs crashes when typing 'memory' into search input

  #6878 opened Jun 26, 2025

• File not supported - @ai-sdk/azure - v5 beta 1

  #6873 opened Jun 26, 2025

• ai-sdk/google error: enum: only allowed for STRING type

  #6872 opened Jun 26, 2025

• Tool parameters zod validator cannot use async refine

  #6868 opened Jun 25, 2025

• Vercel AI SDK returns empty text after MCP tool call resolution

  #6867 opened Jun 25, 2025

• Invalid Tool Parameters when using Anthropic

  #6865 opened Jun 25, 2025

• Streaming doesn't work with tools

  #6861 opened Jun 25, 2025

• Gemini tool parameter descriptions aren't working.

  #6855 opened Jun 24, 2025

• GEMMA-3-12b-it Model Bug with @ai-sdk/google

  #6854 opened Jun 24, 2025

• website breaking on search

  #6852 opened Jun 24, 2025

• Add PDF document support for Anthropic Vertex Provider

  #6830 opened Jun 23, 2025

• toolChoice required doesn't work with vertex gemini-2.0-pro or gemini-2.5-pro

  #6824 opened Jun 22, 2025

• AI SDK v5 Alpha 15 - Stop function (abort signal) with resumable-stream

  #6823 opened Jun 22, 2025

• toolCallStreaming only streaming input, output is not working.

  #6822 opened Jun 22, 2025

• Support global endpoint for Vertex AI models

  #6811 opened Jun 20, 2025

242 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Add support for OpenAI `realtime` voice models.

  #3176 commented on Jun 20, 2025 • 0 new comments

• v5 use-chat-resume example not working

  #6633 commented on Jun 20, 2025 • 0 new comments

• OpenRouter: some openai compatible model use `reasoning` instead of `reasoning_content` as reasoning field name

  #6226 commented on Jun 23, 2025 • 0 new comments

• Support for elevenlabs Audio Streaming

  #777 commented on Jun 23, 2025 • 0 new comments

• Image Embedding Support

  #4125 commented on Jun 25, 2025 • 0 new comments

• Error: ToolInvocation must have a result when processing tool UI parts from streamText

  #6795 commented on Jun 25, 2025 • 0 new comments

• createAzure Does Not Pass Model to Langfuse

  #6777 commented on Jun 25, 2025 • 0 new comments

• `onFinish` does not get called in `streamText` when request stopped by the user from client side using `stop()` function

  #6699 commented on Jun 25, 2025 • 0 new comments

• useObject not streaming text in React Native?

  #6675 commented on Jun 25, 2025 • 0 new comments

• consumeStream doesn't work to prevent client disconnects from cancelling stream from finishing

  #6632 commented on Jun 25, 2025 • 0 new comments

• Add OTEL span attributes for provider metadata

  #6672 commented on Jun 25, 2025 • 0 new comments

• Use more OTEL semantic conventions with ai OTEL spans

  #6673 commented on Jun 25, 2025 • 0 new comments

• incompatible `UIMessage` for different providers

  #6626 commented on Jun 25, 2025 • 0 new comments

• Expo 53 streaming throwing "Can't Find variable: TextEncoderStream"

  #6608 commented on Jun 25, 2025 • 0 new comments

• streamText() does not normalize usage fields from snake_case

  #6564 commented on Jun 25, 2025 • 0 new comments

• Opentelemetry backends fail because of NaN token-count attributes are serialized as `doubleValue: null`

  #6546 commented on Jun 25, 2025 • 0 new comments

• prepareStep should allow message modification

  #6615 commented on Jun 25, 2025 • 0 new comments

• Attempt to log partial object from `streamObject` raises `TypeError`

  #6509 commented on Jun 25, 2025 • 0 new comments

• `experimental_resume` chats don't get stopped by useChat `stop()` / abortSignal

  #6502 commented on Jun 25, 2025 • 0 new comments

• JSON Schema: 'definitions' are not supported with output mode 'array' (OpenAI).

  #6454 commented on Jun 25, 2025 • 0 new comments

• type UseChatHelpers from @ai-sdk/react does not expose addToolResult type #6445

  #6446 commented on Jun 25, 2025 • 0 new comments

• experimental_toToolResultContent not working with image_urls

  #5234 commented on Jun 25, 2025 • 0 new comments

• Angular support?

  #5083 commented on Jun 26, 2025 • 0 new comments

• expose function for processing ui message stream

  #6415 commented on Jun 26, 2025 • 0 new comments

• groq-transcription-model.ts: Transcription Fails Due to Missing File Extension in formData

  #6413 commented on Jun 26, 2025 • 0 new comments

• Optional zod arg is required

  #6403 commented on Jun 26, 2025 • 0 new comments

• Website is scrolling horizontally

  #6390 commented on Jun 26, 2025 • 0 new comments

• Build warning from circular dependency in `zod-to-json-schema`

  #6334 commented on Jun 26, 2025 • 0 new comments

• `headers` option cannot be set without type error in `result.toDataStreamResponse({})` for `streamText()`

  #6253 commented on Jun 26, 2025 • 0 new comments

• openai responses `background` parameter

  #6723 commented on Jun 26, 2025 • 0 new comments

• Maximum update depth exceeded

  #6166 commented on Jun 26, 2025 • 0 new comments

• ai/rsc components streamed through createStreamableUI are mounted multiple times when updating the ui node (update/done)

  #1257 commented on Jun 26, 2025 • 0 new comments

• Examples for function composition

  #1566 commented on Jun 26, 2025 • 0 new comments

• Add Azure Extensions to the Azure provider

  #2237 commented on Jun 26, 2025 • 0 new comments

• Support Azure AI Model Inference API provider

  #2395 commented on Jun 26, 2025 • 0 new comments

• Ungraceful failure error when the anthropic api is overloaded

  #2453 commented on Jun 26, 2025 • 0 new comments

• Support fileId for ImagePart

  #2479 commented on Jun 26, 2025 • 0 new comments

• Add support for moderations

  #2521 commented on Jun 26, 2025 • 0 new comments

• ability to define fallback model / provider

  #2636 commented on Jun 26, 2025 • 0 new comments

• GoogleGenerativeAI crashes when returning value from tool call

  #2711 commented on Jun 26, 2025 • 0 new comments

• Error calling tools on Gemini

  #2926 commented on Jun 26, 2025 • 0 new comments

• Azure OpenAI On Your Data

  #2946 commented on Jun 26, 2025 • 0 new comments

• AWS Bedrock provider doesn't work unless using `bedrockOptions`

  #3018 commented on Jun 26, 2025 • 0 new comments

• More granular control over supported 3rd party API provider features

  #3040 commented on Jun 26, 2025 • 0 new comments

• Support azure authentication with managed identity

  #3124 commented on Jun 26, 2025 • 0 new comments

• support google genai code execution tool

  #3205 commented on Jun 26, 2025 • 0 new comments

• Node.js stream to CLI app using streamText jams up

  #3198 commented on Jun 26, 2025 • 0 new comments

• Gemini Context caching

  #3212 commented on Jun 26, 2025 • 0 new comments

• Allow AI SDK usage without requiring Zod Installation

  #3282 commented on Jun 26, 2025 • 0 new comments

• Azure o1 models: 'max_tokens' is not supported with this model

  #3309 commented on Jun 26, 2025 • 0 new comments

• Google Gen AI: Context caching breaks with generateObject / generateContent -

  #3333 commented on Jun 26, 2025 • 0 new comments

• Amazon Bedrock Provider docs recommend overly permissive IAM policy

  #3365 commented on Jun 26, 2025 • 0 new comments

• StreamObject doesn't work with Anthropic

  #3422 commented on Jun 26, 2025 • 0 new comments

• All number of `maxSteps` are always executed with Google Vertex AI even when it leads to repetition

  #3470 commented on Jun 26, 2025 • 0 new comments

• Getting Error with Azure open ai sdk with retries. Unable to catch the error.

  #3655 commented on Jun 26, 2025 • 0 new comments

• Support Anthropic prompt caching for tool definitions

  #3820 commented on Jun 26, 2025 • 0 new comments

• Azure prompt caching metadata

  #3452 commented on Jun 26, 2025 • 0 new comments

• Multimodal Live API by Google (Gemini)

  #4082 commented on Jun 26, 2025 • 0 new comments

• Support Amazon's new Nova models on bedrock

  #4087 commented on Jun 26, 2025 • 0 new comments

• Support Image generation for @ai-sdk/azure provider

  #4109 commented on Jun 26, 2025 • 0 new comments

• AI assistant needs to be able to handle file attachments as well.

  #4145 commented on Jun 26, 2025 • 0 new comments

• Telemetry: support opentelemetry genai standard prompts in openai format

  #4160 commented on Jun 26, 2025 • 0 new comments

• OpenAI Usage Error break stream

  #4190 commented on Jun 26, 2025 • 0 new comments

• Telemetry: Filter out large data parts

  #4161 commented on Jun 26, 2025 • 0 new comments

• Support "Grounding with your own data" in Google provider

  #4230 commented on Jun 26, 2025 • 0 new comments

• Support for the `promptFeedback.blockReason` in the Gemini streaming output

  #4236 commented on Jun 26, 2025 • 0 new comments

• Imagen 3.0 on Vertex sometimes drops error

  #4302 commented on Jun 26, 2025 • 0 new comments

• "useSearchGrounding" with Google Gen AI breaks tool calling

  #4312 commented on Jun 26, 2025 • 0 new comments

• add character count to telemetry stats

  #4317 commented on Jun 26, 2025 • 0 new comments

• Bedrock produces "ValidationException: This model doesn't support the toolConfig.toolChoice.tool field." when using Nova model.

  #4396 commented on Jun 26, 2025 • 0 new comments

• Add support for Amazon Bedrock prompt caching

  #4472 commented on Jun 26, 2025 • 0 new comments

• Support Anthropic’s Citation API

  #4497 commented on Jun 26, 2025 • 0 new comments

• Using openai o*-mini generates an invalid tool's parameters schema.

  #4662 commented on Jun 26, 2025 • 0 new comments

• Google model content caching stopped working since yesterday

  #4664 commented on Jun 26, 2025 • 0 new comments

• video input in google sdk and gemini template

  #4678 commented on Jun 26, 2025 • 0 new comments

• Support for latest anthropic computer-use version

  #6471 commented on Jun 26, 2025 • 0 new comments

• Enable tracing for image generation

  #4704 commented on Jun 26, 2025 • 0 new comments

• metadataExtractor is not recognized in OpenAICompatibleProviderSettings

  #4715 commented on Jun 26, 2025 • 0 new comments

• must be specified when not using one_of using zod.union() on google provider

  #4725 commented on Jun 26, 2025 • 0 new comments

• Support for Perplexity Sonar Pro Reasoning

  #4751 commented on Jun 26, 2025 • 0 new comments

• File paths with spaces fail when using files with GCP, Vertex and Gemini

  #4759 commented on Jun 26, 2025 • 0 new comments

• Support for Google RetrievalTool in tools options

  #4779 commented on Jun 26, 2025 • 0 new comments

• ai-sdk/azure calling the wrong url path for r1 deployment

  #4797 commented on Jun 26, 2025 • 0 new comments

• CreateGroq Usage Streaming

  #4827 commented on Jun 26, 2025 • 0 new comments

• Add Gemini API using Vertex AI in Firebase

  #4887 commented on Jun 26, 2025 • 0 new comments

• An Error is occurred in tool calls

  #4893 commented on Jun 26, 2025 • 0 new comments

• Replicate language model support

  #4918 commented on Jun 26, 2025 • 0 new comments

• feature-request: expose all chat messages in react hook `useChat.onFinish`

  #4919 commented on Jun 26, 2025 • 0 new comments

• Add converter from type `CoreMessage` to type `Message`

  #4875 commented on Jun 26, 2025 • 0 new comments

• Dynamic retrieval with Gemini 2.0 Flash

  #4946 commented on Jun 26, 2025 • 0 new comments

• Ability to work with Vertex AI Agents

  #4972 commented on Jun 26, 2025 • 0 new comments

• Support for Custom URI Schemes (e.g., s3://) in experimental_attachments

  #4966 commented on Jun 26, 2025 • 0 new comments

• Problems upgrading from Bedrock v1 to v2

  #4996 commented on Jun 26, 2025 • 0 new comments

• Improved Google Generative AI SDK Error Reporting

  #4997 commented on Jun 26, 2025 • 0 new comments

• Support for audio outputs with gpt-4o-audio-preview

  #5007 commented on Jun 26, 2025 • 0 new comments

• Respect `retry-after` header for API (Anthropic at least)

  #5018 commented on Jun 26, 2025 • 0 new comments

• Cohere embeddings models do not work in @ai-sdk/amazon-bedrock

  #5055 commented on Jun 26, 2025 • 0 new comments

• @ai-sdk/google 400 when using nullable/optional fields

  #5057 commented on Jun 26, 2025 • 0 new comments

• Allow passing tokens to embedding functions (instead of / in addition to strings)

  #5067 commented on Jun 26, 2025 • 0 new comments

• Error: Model {modelName} is enabled only for api versions 2024-12-01-preview and later | Azure

  #5106 commented on Jun 26, 2025 • 0 new comments

• collect client-side telemetry (useChat)

  #5114 commented on Jun 26, 2025 • 0 new comments

• Export standardizePrompt

  #5090 commented on Jun 26, 2025 • 0 new comments

• Getting 'Unsupported URL protocol: gs:' when sending a CoreMessage of type file with a GCS URL to streamText using vertex-ai provider

  #5121 commented on Jun 26, 2025 • 0 new comments

• Bedrock request metadata is undefined in streamText() OnFinishResult and Returns

  #5129 commented on Jun 26, 2025 • 0 new comments

• Support for telemetry in image processing

  #5137 commented on Jun 26, 2025 • 0 new comments

• Add system prompt to default settings middleware

  #5186 commented on Jun 26, 2025 • 0 new comments

• Support OpenAI Vector Store with the new Response API

  #5188 commented on Jun 26, 2025 • 0 new comments

• Support for OpenAI Responses status

  #5189 commented on Jun 26, 2025 • 0 new comments

• generateObject for openAICompatible models

  #5197 commented on Jun 26, 2025 • 0 new comments

• createDataStreamResponse doesn't allow streamText to be aborted via signals/useChat.stop method

  #5204 commented on Jun 26, 2025 • 0 new comments

• anthropic count tokens api

  #5205 commented on Jun 26, 2025 • 0 new comments

• Unable to build file: 'this' keyword is equivalent to 'undefined' at the top level of an ES module

  #5210 commented on Jun 26, 2025 • 0 new comments

• Handle Gemini overloaded error

  #5217 commented on Jun 26, 2025 • 0 new comments

• x-ratelimit header

  #5223 commented on Jun 26, 2025 • 0 new comments

• Anthropic Caching is not working with generateObject

  #5227 commented on Jun 26, 2025 • 0 new comments

• Support for provider-defined tools with Amazon Bedrock

  #5266 commented on Jun 26, 2025 • 0 new comments

• Add Reasoning Tokens to Perplexity Provider Metadata

  #5406 commented on Jun 26, 2025 • 0 new comments

• Error: "error" parts expect a string value.

  #5407 commented on Jun 26, 2025 • 0 new comments

• Observation mismatch when running parallel LLM calls using ai and langfuse-vercel

  #5414 commented on Jun 26, 2025 • 0 new comments

• Add support for logprobs in Google's Gemini 2.5/ 2.0 models

  #5418 commented on Jun 26, 2025 • 0 new comments

• Google Vertex Anthropic Image URL Support

  #5449 commented on Jun 26, 2025 • 0 new comments

• Azure OpenAI Responses API (Preview) support

  #5454 commented on Jun 26, 2025 • 0 new comments

• Mistral codestral FIM model endpoint

  #5469 commented on Jun 26, 2025 • 0 new comments

• Support DocX in file attachments

  #5486 commented on Jun 26, 2025 • 0 new comments

• Cache Control in tools with mcp servers

  #5497 commented on Jun 26, 2025 • 0 new comments

• Support file id in openai responses messages

  #5514 commented on Jun 26, 2025 • 0 new comments

• Support parallel tool calls on client

  #5516 commented on Jun 26, 2025 • 0 new comments

• defaulting to an empty object for tools with no parameters

  #5527 commented on Jun 26, 2025 • 0 new comments

• AI_APICallError: Invalid value for 'content': expected a string, got null.

  #5535 commented on Jun 26, 2025 • 0 new comments

• `toolCallStreaming` is not functioning as expected with Bedrock or Google models.

  #5544 commented on Jun 26, 2025 • 0 new comments

• Support for Specifying Task Types in @ai-sdk/google Text Embedding Models

  #5554 commented on Jun 26, 2025 • 0 new comments

• Prioritize User-Provided Image mimeType Over Detected One

  #5575 commented on Jun 26, 2025 • 0 new comments

• Azure - Provider metadata incorrect when using content-filter annotations

  #5577 commented on Jun 26, 2025 • 0 new comments

• useChat throttle might cause the onFinish callback to be triggered at the wrong time

  #5584 commented on Jun 26, 2025 • 0 new comments

• Azure OpenAI ResponsesAPI not usable with tools

  #5592 commented on Jun 26, 2025 • 0 new comments

• streamObject doesn't work with google Gemini 2.5 Pro

  #5594 commented on Jun 26, 2025 • 0 new comments

• streamText onChunk processing ends before the callback promise is resolved

  #5599 commented on Jun 26, 2025 • 0 new comments

• amazon bedrock package incorrectly extracting mimetype from files

  #5606 commented on Jun 26, 2025 • 0 new comments

• experimental_output can be caught in "unending stream of whitespace" responses with OpenAI

  #5608 commented on Jun 26, 2025 • 0 new comments

• export `generateObject` parameter `options` type

  #5615 commented on Jun 26, 2025 • 0 new comments

• Requests to vertex ai fail if `z.record(z.string())` used in tool schema

  #5628 commented on Jun 26, 2025 • 0 new comments

• Missing tools Property in AzureOpenAIProvider Interface

  #5635 commented on Jun 26, 2025 • 0 new comments

• Support for AWS Bedrock Application Inference Profiles (AIP) in @vercel/ai

  #5648 commented on Jun 26, 2025 • 0 new comments

• @ai-sdk/openai streamText can't use tools/functions with no arguments.

  #5658 commented on Jun 26, 2025 • 0 new comments

• MCP Resources integration

  #5702 commented on Jun 26, 2025 • 0 new comments

• Google Gemini model fails to use tools properly in follow-up messages (gemini-2.5-pro-preview-03-25)

  #5717 commented on Jun 26, 2025 • 0 new comments

• Bug: File names in experimental_attachments are replaced with auto-generated IDs when using Bedrock provider

  #5740 commented on Jun 26, 2025 • 0 new comments

• Can't extract type for valid providers from provider registry

  #5745 commented on Jun 26, 2025 • 0 new comments

• Request for an example of how to work with google's new Agent Developer Kit (ADK)

  #5754 commented on Jun 26, 2025 • 0 new comments

• Add Deno example

  #5760 commented on Jun 26, 2025 • 0 new comments

• Using xAI grok3 beta and tools, if a character like 0:"<" appears, it will fail if it fails if it appears

  #5761 commented on Jun 26, 2025 • 0 new comments

• max completion/output token calculations

  #5792 commented on Jun 26, 2025 • 0 new comments

• multi-step tool calls broken with `cohere provider`

  #5794 commented on Jun 26, 2025 • 0 new comments

• Page refreshes itself when submit() is called when using useObject()

  #5795 commented on Jun 26, 2025 • 0 new comments

• Inconsistent Error Handling: Vertex AI on Edge Runtime Returns Empty Error ({}) and Skips Retries, Unlike Standard Google API

  #5797 commented on Jun 26, 2025 • 0 new comments

• Error in docs: Example: Use the re-ask strategy for repair

  #5801 commented on Jun 26, 2025 • 0 new comments

• Tool calls loop infinitely when using streamText with structuredOutputs enabled

  #5875 commented on Jun 26, 2025 • 0 new comments

• allow passing a body object to useObject

  #5877 commented on Jun 26, 2025 • 0 new comments

• Support propertyOrdering for google models

  #5879 commented on Jun 26, 2025 • 0 new comments

• Functional calling is not working from meta/llama-3.3-70b from sglang framework

  #5885 commented on Jun 26, 2025 • 0 new comments

• You must be logged in to use this feature error

  #5903 commented on Jun 26, 2025 • 0 new comments

• Flexible sources interface

  #5939 commented on Jun 26, 2025 • 0 new comments

• providerOptions not applied when using parts in messages

  #5942 commented on Jun 26, 2025 • 0 new comments

• Support image editing with generateImage

  #5949 commented on Jun 26, 2025 • 0 new comments

• Mock functions given in testing part of documentation does not work with edge runtime.

  #5973 commented on Jun 26, 2025 • 0 new comments

• Build failed sveltekit-openai#build

  #6006 commented on Jun 26, 2025 • 0 new comments

• Structured clone used for OTEL fails when nesting spans

  #6030 commented on Jun 26, 2025 • 0 new comments

• Feature request: add `onUpdate` option for `useChat`

  #6054 commented on Jun 26, 2025 • 0 new comments

• handle stream text -> data in the client sdks

  #6060 commented on Jun 26, 2025 • 0 new comments

• Support fine-tuned Gemini models

  #6084 commented on Jun 26, 2025 • 0 new comments

• Feature request: `onToolResult` callback in `useChat`

  #6080 commented on Jun 26, 2025 • 0 new comments

• gemini-2.5-pro-preview-03-25 using Generative AI provider doesn't include thinking tokens in token counts

  #6082 commented on Jun 26, 2025 • 0 new comments

• calculateLanguageModelUsage does not add thoughtTokens from Gemini API responses

  #6089 commented on Jun 26, 2025 • 0 new comments

• Server-Side Utility for Processing Data Streams

  #6090 commented on Jun 26, 2025 • 0 new comments

• after addToolResult, the ai-sdk call /chat twice

  #6123 commented on Jun 26, 2025 • 0 new comments

• Feature request: provide access to OpenAI reasoning tokens when streaming object

  #6126 commented on Jun 26, 2025 • 0 new comments

• is there an api to get the specific Provider's all options

  #6138 commented on Jun 26, 2025 • 0 new comments

• Add thinking support for useCompletition react api

  #6141 commented on Jun 26, 2025 • 0 new comments

• Urgent Guidance Needed: Server `onFinish` lacks consistent Message ID access between `streamText` & `streamObject`

  #6144 commented on Jun 26, 2025 • 0 new comments

• Simulated readable stream does not work with `streamObject` using the array output mode

  #6148 commented on Jun 26, 2025 • 0 new comments

• Google Vertex Imagen Fails To Generate Images

  #6177 commented on Jun 26, 2025 • 0 new comments

• The attachment url is not passed to the LLM

  #6204 commented on Jun 26, 2025 • 0 new comments

• Feature request: Adding prioritization to prompts

  #6246 commented on Jun 26, 2025 • 0 new comments

• Amazon Bedrock + streamText + maxSteps often leads to APICallError

  #6254 commented on Jun 26, 2025 • 0 new comments

• generateText(): finishReason is tool-calls, but toolCalls is empty

  #6257 commented on Jun 26, 2025 • 0 new comments

• Add support for tools to directly return tool results to the client

  #6258 commented on Jun 26, 2025 • 0 new comments

• embedMany not splitting request into proper chunk size for google's text-embedding-004 model

  #6268 commented on Jun 26, 2025 • 0 new comments

• Image generation response missing usage information

  #6275 commented on Jun 26, 2025 • 0 new comments

• MCP Client does not respond to ping requests from server

  #6282 commented on Jun 26, 2025 • 0 new comments

• Add Support for Prompt and Resource Actions in AI SDK for MCP Servers

  #6294 commented on Jun 26, 2025 • 0 new comments

• Azure AI Foundry Support

  #6295 commented on Jun 26, 2025 • 0 new comments

• Vertex Gemini Web Sources not returning

  #6306 commented on Jun 26, 2025 • 0 new comments

• _TypeValidationError, response does not adhere to zod schema

  #6336 commented on Jun 26, 2025 • 0 new comments

• Support access of the `deploymentName` in the Azure OpenAI Provider Language Model

  #6378 commented on Jun 26, 2025 • 0 new comments

• Chat Store Body Fixed to Initial Value

  #6386 commented on Jun 26, 2025 • 0 new comments

• Add imagen 3 to `google.image()`

  #6408 commented on Jun 26, 2025 • 0 new comments

• Support for reasoning model with custom deployment name

  #6410 commented on Jun 26, 2025 • 0 new comments

• generateText(): generated text is empty when used with custom tool

  #6414 commented on Jun 26, 2025 • 0 new comments

• Support for OpenAI's Responses API Image Generation Tool

  #6425 commented on Jun 26, 2025 • 0 new comments

• Thinking Summaries - Gemini 2.5 pro

  #6434 commented on Jun 26, 2025 • 0 new comments

• messages.0.content.0.image.source.base64.data: URL sources are not supported

  #6453 commented on Jun 26, 2025 • 0 new comments

• OpenAI Responses MCP

  #6458 commented on Jun 26, 2025 • 0 new comments

• chatStore error with anthropic-beta: interleaved-thinking-2025-05-14 header set

  #6459 commented on Jun 26, 2025 • 0 new comments

• Request an update for the FASTAPI example using AI SDK v5

  #6465 commented on Jun 26, 2025 • 0 new comments

• Out of order text-delta for anthropic claude 3.5 and 3.7 models from bedrock and vertex apis

  #6467 commented on Jun 26, 2025 • 0 new comments

• Support for computer-use models on AWS bedrock

  #6470 commented on Jun 26, 2025 • 0 new comments

• Support encrypted reasoning items in OpenAI responses API

  #6488 commented on Jun 26, 2025 • 0 new comments

• Support for image URLs for claude 4.0

  #6493 commented on Jun 26, 2025 • 0 new comments

• Support JSON schema for Gemini 2.5

  #6494 commented on Jun 26, 2025 • 0 new comments

• tool-calls will silently fail while LLM generate a wrong json when using `streamText` in `openai-compatible` provider

  #6520 commented on Jun 26, 2025 • 0 new comments

• Support new "refusal" stop reason for Claude 4 models

  #6529 commented on Jun 26, 2025 • 0 new comments

• Visibility over retries

  #6531 commented on Jun 26, 2025 • 0 new comments

• event: errors in OpenAI reasoning stream result in finishReason=unknown

  #6534 commented on Jun 26, 2025 • 0 new comments

• Support Gemini "URL context" tool in google provider

  #6535 commented on Jun 26, 2025 • 0 new comments

• @ai-sdk/openai transcription providerOptions timestamp_granularities is broken

  #6561 commented on Jun 26, 2025 • 0 new comments

• @ai-sdk/openai transcription response_format is missing from the providerOptions

  #6562 commented on Jun 26, 2025 • 0 new comments

• Azure OpenAI Does not output reasoning summary

  #6569 commented on Jun 26, 2025 • 0 new comments

• parameters.properties field missing in request body when using Vertex AI with Vercel AI SDK

  #6572 commented on Jun 26, 2025 • 0 new comments

• Typing error on generateImage

  #6586 commented on Jun 26, 2025 • 0 new comments

• Reasoning is not working - Amazon Bedrock Anthropic

  #6617 commented on Jun 26, 2025 • 0 new comments

• generateText structuredOutputs Zod schema not respecting z.number.multipleOf() or step (potentially OpenAI issue)

  #6642 commented on Jun 26, 2025 • 0 new comments

• `items` can be missing property of `array` when fetching `tools` of an MCP server via `MCPClient`

  #6674 commented on Jun 26, 2025 • 0 new comments

• Tool Calls Fail with Empty Arguments in OpenAI-Compatible Provider

  #6687 commented on Jun 26, 2025 • 0 new comments

• Anthropic/OpenAI Batch API support

  #6701 commented on Jun 26, 2025 • 0 new comments

• Shopify MCP error

  #6704 commented on Jun 26, 2025 • 0 new comments

• add OAuth support to MCP Client

  #6717 commented on Jun 26, 2025 • 0 new comments

• OpenAI-compatible adapter fails to properly handle client tools?

  #6719 commented on Jun 26, 2025 • 0 new comments

• Disable thinking while using OpenAI-compatible chat API interface

  #6728 commented on Jun 26, 2025 • 0 new comments

• Add support for MCP tool annotations

  #6732 commented on Jun 26, 2025 • 0 new comments

• Export Zod schema for UIMessage

  #6763 commented on Jun 26, 2025 • 0 new comments

• `providerOptions.stream_options` No Longer Effective in Latest SDK, Preventing Usage Data for DeepSeek/Volcengine

  #6774 commented on Jun 26, 2025 • 0 new comments

• experimental_repairToolCall is not being called when the toolCall json has some errors

  #6800 commented on Jun 26, 2025 • 0 new comments

• multi_tool_use error but on Azure open ai

  #4085 commented on Jun 26, 2025 • 0 new comments

• openai-compatible can't use for gemini openai endpoint(v1beta/openai),Exception at validation

  #6663 commented on Jun 26, 2025 • 0 new comments

• `streamText` tool invocation does not return the result for Node v20.10.0

  #3263 commented on Jun 26, 2025 • 0 new comments

• Support for reasoning with generateObject

  #6474 commented on Jun 26, 2025 • 0 new comments

• open ai provider tool schema params fails with optional params

  #6776 commented on Jun 26, 2025 • 0 new comments

• The delayInMs parameter supports passing a function.

  #6130 commented on Jun 26, 2025 • 0 new comments

• fix (provider/amazon-bedrock): better resolve mime-types of document and images

  #6185 commented on Jun 25, 2025 • 0 new comments

• feat (provider/google): Added URL context Tool

  #6443 commented on Jun 23, 2025 • 0 new comments

• feat(provider/anthropic): implement web search tool

  #6667 commented on Jun 24, 2025 • 0 new comments

• fix(ai): Azure stream text/object otel fix

  #6803 commented on Jun 24, 2025 • 0 new comments