Migrate AI SDK V6 To V7

AI SDK 6 to 7 Migration

Use content/docs/08-migration-guides/23-migration-guide-7-0.mdx from the AI SDK repo as the source of truth. This skill is the working checklist; read the guide for exact examples or when behavior is unclear.

Migration Workflow

1. Ensure the user has a clean backup or committed baseline before editing.
2. Inspect package.json and lockfiles to identify installed ai, @ai-sdk/*, provider, UI, MCP, and telemetry packages.
3. Upgrade AI SDK packages to latest versions, and add @ai-sdk/otel only if the project uses OpenTelemetry spans.
4. Update runtime and module assumptions: Node.js must be >=22, and AI SDK packages are ESM-only. Replace require() imports with ESM imports and add "type": "module" or use .mjs where needed.
5. Search for the v6 patterns below, migrate only the code that exists, then run typecheck and targeted tests.

Prefer behavior-preserving changes. When v7 changes semantics, decide whether the app wants the new all-steps behavior or the previous final-step-only behavior.

Core API Changes

• experimental_customProvider -> customProvider.
• experimental_generateImage -> generateImage; Experimental_GenerateImageResult -> GenerateImageResult.
• experimental_transcribe -> transcribe; Experimental_TranscriptionResult -> TranscriptionResult.
• experimental_generateSpeech -> generateSpeech; Experimental_SpeechResult -> SpeechResult.
• experimental_output option/result -> output option/result.
• CallSettings -> LanguageModelCallOptions & Omit<RequestOptions, 'timeout'>; prepareCallSettings -> prepareLanguageModelCallOptions.
• stepCountIs -> isStepCount.

Prompts and Steps

• Rename top-level system to instructions for generateText, streamText, generateObject, streamObject, and streamUI.
• Move { role: 'system' } messages from prompt or messages into top-level instructions. Only use allowSystemInMessages: true for trusted persisted messages.
• Rename experimental_prepareStep to prepareStep.
• In prepareStep, rename returned system to instructions.
• In experimental_repairToolCall, use { instructions } instead of { system }.
• Audit prepareStep behavior: returned instructions and messages now carry forward into later steps. If code depended on one-step-only overrides, rebuild from initialInstructions, initialMessages, and responseMessages explicitly.

Lifecycle Callbacks

• experimental_onStart -> onStart.
• experimental_onStepStart -> onStepStart.
• onFinish -> onEnd.
• onStepFinish -> onStepEnd.
• For embed, embedMany, and rerank, experimental_onFinish -> onEnd.
• Callback event fields use instructions instead of system.

Usage, Telemetry, and Include Options

• usage.cachedInputTokens -> usage.inputTokenDetails.cacheReadTokens.
• usage.reasoningTokens -> usage.outputTokenDetails.reasoningTokens.
• OpenTelemetry moved out of ai; install @ai-sdk/otel and call registerTelemetry(new OpenTelemetry(...)) at app startup.
• Telemetry is enabled by default once an integration is registered. Remove redundant isEnabled: true; use isEnabled: false to opt out per call.
• Move experimental_telemetry.tracer into the OpenTelemetry constructor.
• experimental_telemetry -> telemetry.
• Telemetry integration callbacks: onRerankFinish -> onRerankEnd, onEmbedFinish -> onEmbedEnd. Update tracing-channel subscribers for the same event type names.
• experimental_include -> include.
• includeRawChunks -> include.rawChunks.
• Request and response bodies are excluded by default. If code reads request.body or response.body, opt in with include.requestBody and, for generateText, include.responseBody.

Streaming, Messages, and Tools

• StreamTextResult.fullStream -> stream.
• streamText onChunk now receives all stream parts, including lifecycle, boundary, finish, abort, and error parts. Guard by chunk.type before assuming text/tool/raw content.
• step.response.messages is no longer accumulated across previous steps. Use result.responseMessages for the full response message history, or flatten result.steps.
• Tool execution callbacks: experimental_onToolCallStart -> onToolExecutionStart, experimental_onToolCallFinish -> onToolExecutionEnd.
• Tool callback experimental_context -> context.
• Split shared runtime data from tool-specific data: use top-level runtimeContext for orchestration state, declare per-tool contextSchema, and pass per-tool values through toolsContext.
• Move needsApproval from tool() / dynamicTool() into per-call or agent toolApproval.
• experimental_activeTools -> activeTools.
• ToolCallOptions -> ToolExecutionOptions.
• isToolOrDynamicToolUIPart -> isToolUIPart.

Content Parts and Reasoning

• Tool result { type: 'media' } is removed; use { type: 'file-data' }.
• Migrate toModelOutput image-*, file-*, file-id, and image-file-id variants to canonical { type: 'file', mediaType, data: { type: 'data' | 'url' | 'reference', ... } }.
• User message { type: 'image', image, mediaType? } is deprecated; use { type: 'file', mediaType: 'image' | 'image/*', data }.
• Add support for the new reasoning-file content type in exhaustive switches, renderers, serializers, and validators.
• When adopting top-level reasoning, remove overlapping provider-specific reasoning settings from providerOptions unless provider-specific settings intentionally take precedence.

Multi-Step Result Shape

• result.usage now includes all steps; result.totalUsage is deprecated. Use result.finalStep.usage for final-step-only usage.
• Top-level content, toolCalls, staticToolCalls, dynamicToolCalls, toolResults, staticToolResults, dynamicToolResults, files, sources, and warnings now include all steps. Use finalStep for previous final-step-only behavior.
• Top-level reasoning, reasoningText, request, response, and providerMetadata are deprecated for final-step data. Use result.finalStep.*; for streamText, await result.finalStep.
• Apply the same result-shape rules to onEnd events.

Stream Response Helpers

The streamText result helper methods are deprecated. Replace result methods with top-level stateless helpers:

• result.toUIMessageStream(...) -> toUIMessageStream({ stream: result.stream, ... }).
• result.toUIMessageStreamResponse(...) -> toUIMessageStream(...) plus createUIMessageStreamResponse({ stream }).
• result.pipeUIMessageStreamToResponse(response, ...) -> toUIMessageStream(...) plus pipeUIMessageStreamToResponse({ response, stream }).
• result.toTextStreamResponse() -> toTextStream({ stream: result.stream }) plus createTextStreamResponse({ stream }).
• result.pipeTextStreamToResponse(response) -> toTextStream({ stream: result.stream }) plus pipeTextStreamToResponse({ response, stream }).

Package-Specific Checks

• MCP: MCPTransportConfig.redirect now defaults to 'error'. Only set redirect: 'follow' for trusted MCP servers that rely on redirects.
• Vue: @ai-sdk/vue Chat class is deprecated. Prefer useChat, including getter/ref init for reactive chat inputs.
• Anthropic and @ai-sdk/google-vertex/anthropic: providerMetadata.anthropic.cacheCreationInputTokens was removed. Use usage.inputTokenDetails.cacheWriteTokens; raw Anthropic usage remains at finalStep.providerMetadata?.anthropic?.usage.
• Google: rename GoogleGenerativeAI* types, classes, and functions to Google*, e.g. createGoogleGenerativeAI -> createGoogle. The google entry point is unchanged.

Validation

Run the project typecheck after edits, then the smallest relevant test suite. Also smoke-test streaming, chat UI, tool execution, telemetry, and multi-step flows if the migration touched them. If type errors remain, search the migration guide for the exact removed or renamed symbol before inventing a workaround.