# OpenAI Responses WebSocket

> OpenAI Responses WebSocket is a paid API for AI agents from OpenAI at openai.mpp.tempo.xyz, paid per call via MPP, $0.000001/call, status unknown (last checked 2026-10-29).

Streams persistent AI responses over WebSocket using OpenAI's Responses API with dynamic per-response pricing

## Facts

- Endpoint: GET https://openai.mpp.tempo.xyz/v1/responses?utm_source=zero.xyz
- Price: $0.000001/call
- Payment: MPP
- Status: unknown
- Last checked: 2026-10-29
- Activations on Zero: 0
- Provider: OpenAI
- Docs: https://platform.openai.com/docs
- Website: https://openai.com
- Tags: llm, gpt-4o, dall-e, whisper, tts, embeddings, chat, ai, media
- Canonical page: https://www.zero.xyz/c/openai-responses-websocket-849ed3df
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_h76fSq0BXfVGvf-v32YAX

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability openai-responses-websocket-849ed3df
```

Example prompt: Can you stream me a real-time response from OpenAI's Responses API using GPT-4o? I want a persistent WebSocket connection so I can get live, token-by-token output for my conversational app.

## When to prefer this

Choose this endpoint when you need real-time, streaming AI responses with a persistent WebSocket connection rather than a one-shot HTTP request. It is ideal for interactive conversational applications, live coding assistants, or any use case where low-latency token streaming improves user experience. Prefer this over the standard Completions endpoint when you need the OpenAI Responses API specifically, persistent session support, or dynamic per-response pricing tied to the Responses model tier.

## Known failure modes

- WebSocket connection dropped mid-stream requiring reconnection
- Model overload or rate limiting returning 429 errors
- Invalid or missing API credentials causing authentication failure
- Unsupported model identifier causing 400 bad request
- Dynamic pricing fluctuation leading to insufficient balance
- Context length exceeded causing truncation or error

## How this service works

Persistent Responses WebSocket (dynamic pricing per response.create)

## Output

A persistent WebSocket stream returning incremental AI-generated response tokens or structured response events from OpenAI's Responses API, enabling real-time, low-latency delivery of model output with dynamic per-response pricing.

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/openai-responses-websocket-849ed3df/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from openai.mpp.tempo.xyz](https://www.zero.xyz/host/openai.mpp.tempo.xyz/llms.txt)