Sign in
Autofixer

Start an autofixer analysis run for the given error group. Creates the run record immediately and spawns analysis in the background.

post/projects/{project_id}/autofixer/analyze

Request

  • Base URL: /api (relative, and the document was published on a file host, not at the API’s origin)
  • URL: /api/projects/{project_id}/autofixer/analyze
  • Auth: HTTP bearer

Path parameters

project_idinteger required

Project ID

Request body

branchstring nullable

Branch to clone instead of the project's main branch.

error_group_idinteger required
max_turnsinteger nullable

Per-run turn cap applied to every phase (1–200). Only enforced for CLIs with a turn flag (Claude Code). None uses the provider's configured defaults.

modelstring nullable

Model id for the chosen provider. None uses the provider's saved default model.

providerstring nullable

AI provider id ("claude_cli", "codex_cli", "opencode"). None uses the platform default provider.

user_contextstring nullable

Free-text notes for the model (extra context about the error, retry guidance, constraints). Included verbatim in the analysis prompt.

Response

Analysis started; returns run_id for streaming

ai_modelstring nullable
ai_outputstring nullable
ai_providerstring nullable

AI provider slug this run executes with (e.g. claude_cli, codex_cli).

analysisstring nullable
branch_namestring nullable
completed_atstring nullable
created_atstring required
error_messagestring nullable
files_changedinteger required
idinteger required
phasestring nullable
pr_numberinteger nullable
pr_urlstring nullable
project_idinteger required
started_atstring nullable
statusstring required
tokens_inputinteger required
tokens_outputinteger required
trigger_source_idinteger nullable
user_contextstring nullable

Changes

No recorded changes to this operation across all 50 revisions of this API.