Refine a prompt version | API Reference

1 curl -X POST http://localhost:8000/api/prompts/1/versions/1/refine \ 2 -H "Authorization: Token <api_key>" \ 3 -H "Content-Type: application/json" \ 4 -d '{ 5 "teacher_model_provider_connection_id": 1, 6 "teacher_model_name": "teacher_model_name", 7 "project_id": 1 8 }'

1 { 2 "prompt": "prompt", 3 "title": "title", 4 "reasoning": "reasoning", 5 "refinement_job_id": "refinement_job_id", 6 "refinement_status": "refinement_status", 7 "total_cost": "total_cost", 8 "previous_version": { 9 "id": 1, 10 "created_by": { 11 "id": 1, 12 "avatar": "avatar", 13 "first_name": "first_name", 14 "last_name": "last_name", 15 "email": "email" 16 }, 17 "score": "score", 18 "title": "title", 19 "prompt": "prompt", 20 "provider_model_id": "provider_model_id", 21 "created_at": "2024-01-15T09:30:00Z", 22 "updated_at": "2024-01-15T09:30:00Z", 23 "parent_model": 1, 24 "provider": "OpenAI", 25 "model_provider_connection": 1, 26 "organization": 1 27 } 28 }

Refine a prompt version using a teacher model and save the refined prompt as a new version.

Path parameters

prompt_idintegerRequired

version_idintegerRequired

Query parameters

asyncbooleanOptional

Whether to run the refinement asynchronously

Request

This endpoint expects an object.

teacher_model_provider_connection_idintegerRequired

Model Provider Connection ID to use to refine the prompt

teacher_model_namestringRequired>=1 character

Name of the model to use to refine the prompt

project_idintegerRequired

Project ID to target the refined prompt for

Response

Refined prompt response

promptstring

The refined prompt text

titlestring or null

Title of the refined prompt

reasoningstring or null

Reasoning behind the refinement

refinement_job_idstring or null

Unique identifier for the refinement job

refinement_statusstring or null

Status of the refinement job

total_coststring or null

Total cost of the refinement job (in USD)

previous_versionobject or null

Previous version of the prompt

Path parameters

Headers

Query parameters

Request

Response