01 / DECIDE
Architecture decision
Use it for multi-turn products where continuity, deletion, audit, and cross-device behavior must be predictable.
Decide whether the application resends history, chains response identifiers, or uses a conversation resource, and make that choice visible in data policy. Map input, output, state, and side effects as one observable path before deciding which layer owns choose conversation state explicitly.
02 / BUILD
Three-step implementation
- 01
Frame the contract
Write down the caller, data classification, success condition, timeout, cancellation, and ownership. Use it for multi-turn products where continuity, deletion, audit, and cross-device behavior must be predictable.
- 02
Build one narrow path
Implement one end-to-end path with request correlation, typed state, and reversible failure handling. Decide whether the application resends history, chains response identifiers, or uses a conversation resource, and make that choice visible in data policy.
- 03
Prove the outcome
Turn acceptance into a repeatable fixture, contract test, or browser test. Lifecycle tests cover new, continue, fork, expire, delete, export, and unauthorized conversation access.
03 / BOUND
Production boundary
Never assume provider-side state matches your user account, authorization, or legal retention requirements.
04 / PROVE
Acceptance evidence
Lifecycle tests cover new, continue, fork, expire, delete, export, and unauthorized conversation access.
SOURCE / HTTP
Reproducible source probe
curl -fsSI 'https://developers.openai.com/api/docs/guides/conversation-state' | sed -n '1,5p'