openaiResponsesStorageAdapter
Uses OpenAI itself as the state store. Conversation and response calls are forwarded to OpenAI's
/conversations and /responses endpoints. No Postgres, no Prisma — OpenAI owns the state.Only compatible with the Responses surface on OpenAI.
Signature
openaiResponsesStorageAdapter({
deferItemCreationUntilRun?: boolean,
})
deferItemCreationUntilRun — when true, conversation items aren't written to OpenAI until the run is actually created. Useful when you want to validate a request before spending tokens. Defaults to false.Example
import OpenAI from 'openai'
import {
supercompat,
openaiClientAdapter,
openaiResponsesRunAdapter,
openaiResponsesStorageAdapter,
} from 'supercompat/openai'
const client = supercompat({
clientAdapter: openaiClientAdapter({ openai: new OpenAI() }),
storageAdapter: openaiResponsesStorageAdapter(),
runAdapter: openaiResponsesRunAdapter(),
})
const response = await client.responses.create({
model: 'gpt-4.1-mini',
input: 'Hello.',
})
Deferred item creation
openaiResponsesStorageAdapter({ deferItemCreationUntilRun: true })