What Happens on September 9, 2025? Keeping the Standard Voice You Love

TL;DR: OpenAI plans to retire Standard Voice Mode on September 9, 2025. We don’t know the exact UI or voice changes that will land that day, but here’s our commitment: Say, Pi will adapt quickly to keep the same calm, patient presence you’re used to inside ChatGPT. There may be brief interruptions while we ship updates. Our goal is continuity and fidelity — keeping the feel you love and, as much as possible, the voices themselves.

What we know (and don’t)

To the best of our knowledge, OpenAI intends to streamline ChatGPT voice into a single, advanced experience. That could mean:

• No change for Read Aloud and the familiar voice options (best case), or
• Read Aloud and/or the original SVM voices change or disappear (we hope not, but we’re preparing).

We don’t control these switches, and we won’t pretend we do. What we can control is how fast we adapt on your behalf.

Our plan to maintain continuity

Plan A — Keep using Read Aloud (if unchanged).
If ChatGPT’s Read Aloud remains available with the same voices, nothing changes for you. You’ll keep hearing the voice you picked, with the steady cadence you trust.

Plan B — Our own streaming audio (if Read Aloud and/or voices disappear).
If OpenAI removes or alters those voices, Say, Pi will switch to generating audio inside the extension:

• If OpenAI provides an official API path to the original voices, we’ll use (and pay for) those so you can keep them.
• If not, we’ll create high-fidelity continuity versions (clearly labeled “-style,” e.g., Cove-style). These are tuned for negligible perceived difference in tone, pacing, and warmth — as close as we can responsibly make them.
• Only if that’s legally or technically blocked will we fall back to a more general “calm voice” pack.

We won’t promise “seamless.” We will promise speed, transparency, and care. Expect possible short interruptions while we ship updates, and honest notes about what changed.

Why this matters

If you’re here, you already know: the bond many people formed with voices like Cove, Breeze, Vale, and Juniper wasn’t about flashy prosody — it was about presence. Full, steady answers. Gentle pacing. Feeling heard at two in the morning. Our job is to honor that, not with petitions, but with product.

Pricing that keeps this sustainable

• Free: small evaluation minutes so you can try it.
• Premium: monthly minutes (conservatively sized), plus top-up packs for heavy use.
• No bring-your-own-key complexity — this stays consumer-simple, and it lets us fund the minutes required to keep you talking.

We’ll revisit tiers if OpenAI offers official access to the originals (which could lower our costs).

Privacy you don’t have to think about

Your microphone audio is used only to transcribe your speech. It’s sent to our API (and trusted providers strictly for processing), handled transiently, and not retained or used for training. We don’t sell data. We collect only minimal telemetry to keep things running (e.g., minutes used, errors). You can opt out in settings. Full details are in our Privacy Policy.

What you may notice

• A banner the first time we switch engines explaining what changed.
• A voice picker that makes sources clear: Familiar (when available), Continuity (-style), and a “Classic pacing” toggle for the calm delivery many prefer.
• If there’s a short gap during the transition, it will be measured in hours, not weeks. That’s our target.

How you can help

Tell us how close it sounds. If anything feels off — cadence, warmth, interruptions — send a note. We build for you, and your feedback is the fastest way we tune back to perfect.

The promise

We can’t control what happens on September 9. We can control our response. Whatever changes, we’ll move quickly to keep the presence you trust — and, wherever possible, the very voices you love.

Keep the Standard feel. We’re on it.