GPT WORLD / INDEPENDENT CODEX FIELD MANUAL Source review · 2026-07-31

Retrospective edition / VISION

Treat image understanding as evidence extraction

Define what the model may infer from an image, preserve the original asset reference, and separate observation from downstream decisions.

01 / DECIDE

Architecture decision

Use it for document intake, visual QA, accessibility assistance, and workflows that combine text with bounded image evidence.

Define what the model may infer from an image, preserve the original asset reference, and separate observation from downstream decisions. Map input, output, state, and side effects as one observable path before deciding which layer owns treat image understanding as evidence extraction.

02 / BUILD

Three-step implementation

  1. 01

    Frame the contract

    Write down the caller, data classification, success condition, timeout, cancellation, and ownership. Use it for document intake, visual QA, accessibility assistance, and workflows that combine text with bounded image evidence.

  2. 02

    Build one narrow path

    Implement one end-to-end path with request correlation, typed state, and reversible failure handling. Define what the model may infer from an image, preserve the original asset reference, and separate observation from downstream decisions.

  3. 03

    Prove the outcome

    Turn acceptance into a repeatable fixture, contract test, or browser test. A labeled image set measures extraction accuracy, abstention, OCR failures, and sensitivity to crop and resolution.

03 / BOUND

Production boundary

Do not treat visual output as identity verification, medical diagnosis, or a sole basis for consequential decisions.

04 / PROVE

Acceptance evidence

A labeled image set measures extraction accuracy, abstention, OCR failures, and sensitivity to crop and resolution.

SOURCE / HTTP

Reproducible source probe

curl -fsSI 'https://developers.openai.com/api/docs/guides/images-vision' | sed -n '1,5p'