> ## Documentation Index > Fetch the complete documentation index at: https://docs.qa.tech/llms.txt > Use this file to discover all available pages before exploring further. # Prepare change review > Start a GitHub pull request's change review before its preview is deployed. Call it when the PR opens or gets a new commit. Returns 202 with the review conversation, reused the same way as `POST /chat/change-review`. The agent reads the change and creates the tests it plans to run, without running them. Once the preview is deployed, send `POST /chat/change-review` with `applicationOverrides` for the same head commit: QA.tech treats it as the deployment being ready, and the waiting review runs its tests there. Does nothing if the head commit already has a review in the conversation. Poll `get_chat_conversation` until its `status` is `completed`, `failed`, or `cancelled`; a prepared conversation completes once the tests are created. ## OpenAPI ````yaml /api-reference/api.json post /v1/chat/change-review/prepare openapi: 3.0.0 info: title: QA.tech API version: 1.0.0 contact: name: QA.tech url: https://qa.tech email: support@qa.tech description: API for triggering and managing AI-powered test runs servers: - url: https://api.qa.tech description: Production variables: {} security: - BearerAuth: [] tags: - name: Runs - name: Infrastructure - name: Test Cases - name: Application Builds - name: Status badge - name: Remote Tunnels - name: Chat - name: Applications - name: Features - name: Metrics - name: Projects - name: Knowledge - name: Rules - name: Issues - name: Automations paths: /v1/chat/change-review/prepare: post: tags: - Chat summary: Prepare change review description: >- Start a GitHub pull request's change review before its preview is deployed. Call it when the PR opens or gets a new commit. Returns 202 with the review conversation, reused the same way as `POST /chat/change-review`. The agent reads the change and creates the tests it plans to run, without running them. Once the preview is deployed, send `POST /chat/change-review` with `applicationOverrides` for the same head commit: QA.tech treats it as the deployment being ready, and the waiting review runs its tests there. Does nothing if the head commit already has a review in the conversation. Poll `get_chat_conversation` until its `status` is `completed`, `failed`, or `cancelled`; a prepared conversation completes once the tests are created. operationId: PrepareChangeReviewChat parameters: [] requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/PrepareChangeReviewRequest' responses: '202': description: >- The request has been accepted for processing, but processing has not yet completed. content: application/json: schema: $ref: '#/components/schemas/ChatConversationResponse' components: schemas: PrepareChangeReviewRequest: type: object required: - prUrl properties: prUrl: type: string format: uri context: type: string forceNewConversation: type: boolean description: >- Force a brand-new chat instead of reusing the pull request's existing review conversation. Defaults to false, so repeated reviews of the same PR share one chat unless that chat is archived. default: false postMerge: allOf: - $ref: '#/components/schemas/ChangeReviewPostMergeOptions' description: >- Optional post-merge options for this PR. Stored on the review conversation and applied when the PR merges. Overrides GitHub App integration settings for this PR only. triage: allOf: - $ref: '#/components/schemas/ChangeReviewTriageOptions' description: >- Optional PR triage options for this review. GitHub pull requests only. Overrides the GitHub App integration PR triage toggle and prompt for this review. actor: type: string description: 'Username that triggered the CI job. Example: `octocat`.' branch: type: string description: >- Branch the run was triggered from. Example: `main` or `refs/heads/main`. commitHash: type: string description: >- Full commit SHA the run was triggered for. Used to link the run back to the commit. commitMessage: type: string description: Commit message subject for the triggering commit. repository: type: string description: >- Repository the run was triggered from, as `owner/name`. Example: `QAdottech/qatech`. projectShortId: allOf: - $ref: '#/components/schemas/ProjectShortId' description: >- Target a specific project by its prefixed short ID (proj-slug_shortId or proj_shortId), as returned by the projects API. Required for organization-scoped API keys; project-scoped keys may only pass their own project's short ID. description: >- Prepare a change review for a GitHub pull request before its preview is deployed ChatConversationResponse: type: object required: - shortId - url - createdAt - updatedAt - status - runs properties: shortId: $ref: '#/components/schemas/ChatConversationShortId' url: type: string description: Dashboard URL for the conversation. title: type: string createdAt: type: string format: date-time updatedAt: type: string format: date-time source: type: string enum: - api - ui - github - gitlab - system description: >- Where the conversation originated. `api` for conversations created via this API, `ui` for the dashboard, `github`/`gitlab` for VCS-triggered chats, and `system` for everything else. status: type: string enum: - in_progress - completed - failed - cancelled description: >- Overall conversation lifecycle. `in_progress` while the assistant is working or a background test run / research task is still going. Assistant message `COMPLETED` is not enough — starting a run ends the current turn. Poll until this is `completed`, `failed`, or `cancelled`. runs: type: array items: $ref: '#/components/schemas/ChatConversationRun' description: >- Test runs started from this conversation, newest first. Empty until a run is created. Use `shortId` with `get_run` (`testCases` = all) or nested test-case short IDs from that call with `get_run_test_case`. messages: type: array items: $ref: '#/components/schemas/ChatMessageItem' description: >- Recent messages, newest first. Empty on creation; poll `get_chat_conversation` to retrieve. Internal wake messages are omitted. description: Conversation data with recent messages (newest first). ChangeReviewPostMergeOptions: type: object required: - enabled properties: enabled: type: boolean description: >- When true, run the post-merge agent after this PR merges. When false, skip it even if the GitHub App integration has it enabled. prompt: type: string maxLength: 4000 description: >- Goal for the post-merge agent. This prompt is the entire task — the agent only changes tests when the prompt asks it to. Omit or leave empty to use the default maintenance prompt. description: >- Optional post-merge agent options for this pull request. When set, overrides the GitHub App integration post-merge settings for the agent that runs after this PR merges. Omit to use the integration settings. ChangeReviewTriageOptions: type: object properties: enabled: type: boolean description: >- When true, triage may skip this review even if the GitHub App integration has PR triage off. When false, never skip this review, including a commit triage already skipped. Omit to follow the integration toggle. prompt: type: string maxLength: 4000 description: >- PR triage prompt for this review, replacing the GitHub App integration prompt. Describes which changes to skip and which to review. Does not turn triage on by itself. Omit or leave empty to use the integration prompt. description: >- Optional triage options for this GitHub pull request review. Each field overrides the matching GitHub App integration PR triage setting for this review only. Omit a field to use the integration setting. ProjectShortId: type: string pattern: ^proj(-.+_.+|_.+)$ ChatConversationShortId: type: string pattern: ^chat(-.+_.+|_.+)$ ChatConversationRun: type: object required: - shortId - status - result - url properties: shortId: type: string status: type: string description: >- Run lifecycle status: `INITIATED`, `RUNNING`, `COMPLETED`, `ERROR`, or `CANCELLED`. result: type: string nullable: true description: >- Run outcome once complete: `PASSED`, `FAILED`, or `SKIPPED`. `null` while the run is still in progress. url: type: string description: Dashboard URL for the run. description: >- A test run started from this conversation. `shortId` works with `get_run` and `rerun_run`. ChatMessageItem: type: object required: - id - role - createdAt - text properties: id: type: string role: type: string enum: - user - assistant createdAt: type: string format: date-time text: type: string description: The plain-text body of the message. status: type: string enum: - INITIATED - PARTIAL - COMPLETED - CANCELLED - FAILED description: >- Lifecycle status of this message. Do not treat assistant `COMPLETED` as the end of the conversation — starting a test run ends the current turn while the run continues in the background. Poll the conversation `status` instead. isStreaming: type: boolean description: True while the assistant is still streaming tokens for this message. description: >- A chat message. Tool calls and reasoning are not exposed in this version; only the text response is returned. securitySchemes: BearerAuth: type: http scheme: Bearer ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.