Sign in
OpenAI Chat

Create an OpenAI-compatible response

Create a response using Telnyx's OpenAI-compatible Responses API. This endpoint is compatible with the OpenAI Responses API and may be used with the OpenAI JS or Python SDK by setting the base URL to https://api.telnyx.com/v2/ai/openai.

The conversation parameter refers to a Telnyx Conversation rather than an OpenAI-hosted conversation object. To persist a thread across turns, first create a conversation with POST /ai/conversations, then pass that conversation's id in the Responses request as conversation. The endpoint appends the new input, assistant output, reasoning, and tool-call messages to that conversation. Reuse the same conversation id on subsequent Responses requests, including tool-result followups, so the model receives the prior context.

If conversation is omitted, the request is processed without persisting messages to a Telnyx conversation. Use the Conversations API to manage history: list conversations (optionally filtered by metadata), fetch messages for a conversation, and optionally add messages outside the Responses flow.

You can attach arbitrary metadata when creating a conversation (for example to tag the conversation's source, channel, or user) and later filter by it when listing conversations.

post/ai/openai/responses

Request body

modelstring

Model identifier to use for the response, for example zai-org/GLM-5.1-FP8 or another model available from the Telnyx OpenAI-compatible models endpoint.

service_tierstring

The service tier to use for this request. Supported values vary by model; use GET /v2/ai/openai/models and inspect the model's service_tiers field. If omitted, Telnyx-hosted models use default.

region'USA' | 'EU' | 'AUS' | 'UAE'

Optional data-residency region the request should be served from, using the same vocabulary as your account's Data Locality setting. Behavior depends on mode. Supported for Telnyx-hosted models only: a request routed to an external provider never passes through Telnyx model routing, so a region cannot be enforced for it. Omit for today's latency-based routing.

mode'preferred' | 'strict'

How strictly region is applied. preferred (the default when region is set) tries that region first and falls back to another when the model cannot be served there, so a request that would have succeeded still succeeds. strict pins the request: it is served from that region or it fails with a 422, never redirected to another region. Requires region.

{"stackTrail":"paths:/ai/openai/responses:post:requestBody:content:application/json:schema:properties:input","oasType":"schema","type":"unknown","description":"The input items for this turn, using the OpenAI Responses API input format."}
conversationstring uuid

Optional Telnyx Conversation ID from POST /ai/conversations. When provided, Telnyx stores this turn on that conversation and uses the conversation's prior messages as context. Reuse the same ID for subsequent turns and tool-result followups. Omit it for a non-persisted, stateless response.

instructionsstring

Optional system/developer instructions for the model. When used with a persisted conversation, send these on the first request that creates the thread; subsequent turns can rely on the stored history.

streamboolean

Set to true to stream Server-Sent Events, matching OpenAI's Responses streaming format.

Response

Successful Response

object required

Changes

Changed in 2 of the 99 revisions of this API.4

  • a6a61a29ecdd3See the full diff
    • ○

      added the new optional request property

      new-optional-request-property

    • ○

      added the new optional request property

      new-optional-request-property

    • ○

      added the new optional request property

      new-optional-request-property

    This revision also has 34 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog

  • f8c0ade274501See the full diff
    • ○

      added the new optional request property

      new-optional-request-property

Of the 99 revisions, 1 has no diff computed.