feat: support claude agent SDK-style structured outputs in the OpenCode SDK #8161
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add outputFormat option to session.prompt() for requesting structured JSON output. When type is 'json_schema', injects a StructuredOutput tool that validates model output against the provided schema. - Add OutputFormat schema types (text, json_schema) to message-v2.ts - Add structured_output field to AssistantMessage - Add StructuredOutputError for validation failures - Implement createStructuredOutputTool helper in prompt.ts - Integrate structured output into agent loop with retry support - Regenerate OpenAPI spec with new types - Add unit tests for schema validation
Regenerate SDK types to include outputFormat and structured_output fields.
- Fix loop exit condition to check processor.message.finish instead of result === "stop" (processor.process() returns "continue" on normal exit) - Add system prompt instruction when json_schema mode is enabled to ensure model calls the StructuredOutput tool - Add integration tests for structured output functionality - Fix test Session.messages call to use object parameter format
Document the outputFormat feature for requesting structured JSON output: - Basic usage example with JSON schema - Output format types (text, json_schema) - Schema configuration options (type, schema, retryCount) - Error handling for StructuredOutputError - Best practices for using structured output
When structured output is requested, the model must call the StructuredOutput tool instead of responding with plain text. Changes: - Add toolChoice parameter to LLM.StreamInput - Pass toolChoice: "required" to streamText when outputFormat is json_schema - This forces the model to call a tool, ensuring structured output is captured
Merge upstream changes while preserving structured output feature: - Keep tools deprecation notice from upstream - Keep bypassAgentCheck parameter from upstream - Keep variant field on user messages from upstream - Preserve outputFormat and StructuredOutput tool injection Co-Authored-By: Claude Opus 4.5 <[email protected]>
Author
|
cc @thdxr :) |
Contributor
|
The following comment was made by an LLM, it may be inaccurate: No duplicate PRs found |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
This PR adds support for structured generation in the OpenCode SDK (v2) by implementing an injected
StructuredOutputtool with an arbitrary user-defined JSON schema.This implementation approach follows Claude Code's approach as indicated by debugging and proxy analysis.
When set,
toolChoiceis forced to"required"and the agent loop is exited once the input to the tool call passes validation against the user-supplied JSON Schema.Retries are supported, and retry count is customizable (important for smaller models or providers which have poor support for structured generation!).
Prompts and descriptions are based on those used in the Claude Agent SDK
You can learn more about structured outputs in the Claude Agent SDK here
Closes #5639
How did you verify your code works?
This PR contains both unit tests and integration tests.