Disclosure: this page is written by Swift Labs, the maker of TTS Swap, one of the options below. We have tried to keep the facts separate from the pitch. TTS Swap is an independent service by Swift Labs. It is not affiliated with or endorsed by OpenAI. OpenAI and tts-1 are trademarks of OpenAI.

The facts (from OpenAI's deprecations page)

• OpenAI listed the deprecation under "2026-10-01: Text-to-speech models" and says the models "will be removed from the API on January 6, 2027, with at least three months' notice."
• After a shutdown date, OpenAI says the model "will no longer be accessible."
• OpenAI points to its Realtime API guide for migration. For most apps, that means moving off the simple /v1/audio/speech request.
• To check programmatically, OpenAI says to call GET /v1/models and read each model's shutdown_date field.

Source: OpenAI API deprecations, https://platform.openai.com/docs/deprecations (checked November 2, 2026).

Who is affected

Any code that calls audio.speech.create(model="tts-1", ...) (or tts-1-hd / gpt-4o-mini-tts) will stop working on January 6, 2027. Search your repos for tts-1 and audio.speech.

Your options

1. Migrate to gpt-realtime-2.1-mini. This is OpenAI's recommended path. You stay with OpenAI, but your code has to move to the Realtime API.
2. Switch to another hosted TTS provider. You get a new SDK, new voices and new pricing, and you rewrite the integration.
3. Self-host an open-source TTS model. You get full control, and you also run the servers.
4. Use a drop-in /v1/audio/speech replacement such as TTS Swap (ours). Change your base URL and API key; no rewrite if your code accepts WAV.

TTS Swap: what it is, and what it isn't

TTS Swap (https://tts.deadlinkpatrol.com) answers the same /v1/audio/speech request your code sends today:

client = OpenAI(base_url="https://tts.deadlinkpatrol.com/v1", api_key="sk-tts-…")
client.audio.speech.create(model="tts-1", voice="alloy", input="Hello")

• Free key with just an email: 10,000 characters a month, 5 requests per minute, no card. Get a free key.
• $19/month, 1,000,000 characters per month per key. Cancel anytime yourself.
• Same request and error shape, so the official openai Python and Node libraries work.

Limits:

• One voice per language. Audio comes from the open-source MeloTTS model on Cloudflare Workers AI. Every voice name (alloy, nova, onyx…) gets that same voice, and it will not sound like OpenAI's.
• WAV output, even when you ask for mp3. pcm returns raw 24 kHz 16-bit mono. opus, aac and flac return a 400 error.
• speed is accepted but not applied yet.
• English by default. es, fr, zh, ja and ko are available through extra_body={"language": "es"}.
• 4,096 characters per request, 20 requests per minute per key, and a shared daily capacity cap. Requests over a limit get a 429 and aren't charged.

If you need OpenAI's voices, MP3, or speed control, TTS Swap is not the right fit. Use option 1 or 2. If you need your existing code to keep producing speech after January 6 with just a base URL and API key change (no rewrite if your code accepts WAV), it may be.

Source for TTS Swap details: https://tts.deadlinkpatrol.com (checked November 2, 2026).

FAQ

When does tts-1 stop working? On January 6, 2027, according to OpenAI's deprecations page.

Is tts-1-hd affected too? Yes, on the same date. So is gpt-4o-mini-tts.

What does OpenAI recommend instead? gpt-realtime-2.1-mini.

Is there a tts-1 alternative that needs no code change? Any service that serves the /v1/audio/speech API shape works with a base URL and API key change. TTS Swap is one, with the limits listed above; no rewrite if your code accepts WAV.

— Swift Labs. Migration guide: https://tts.deadlinkpatrol.com/migrate/