Description
Asking LLMs for JSON brings malformed output, missing fields and hand-written parsing and retries. Instructor gets reliable structured data from any LLM using Pydantic: define a model and get typed results.
Validation, type safety and IDE support are built in, with no JSON parsing or retry code, across OpenAI and other providers.
Pydantic models:Define output as classes.
Validation:Automatic retries.
Type safety:IDE completion.
Many models:Major providers.
Validation, type safety and IDE support are built in, with no JSON parsing or retry code, across OpenAI and other providers.
Features
Pydantic models:Define output as classes.
Validation:Automatic retries.
Type safety:IDE completion.
Many models:Major providers.
