Support

Chatter is private A2–B1 English speaking practice. Daily Rehearsal guides four spoken turns through one situation. When bounded local checks accept a wording suggestion that preserves your meaning, Chatter shows the changed words and a short reason. Say It Again lets you rehearse the exact checked line, and Practice Again can return an accepted phrase later from the saved source. You can also speak or type in free-form conversations, everyday situations, and workplace roleplays.

Yes. On the subscription screen, tap See how Chatter works. The same walkthrough is available from the locked Practice banner. It demonstrates Say, Check, Say Again, and Return without a purchase, model download, microphone, or saved conversation. It is an authored product walkthrough, not a free AI-practice plan.

Chatter requires iOS 26 or iPadOS 26. It uses Apple's on-device model when that runtime is available. Another supported device may be offered a compatible local-model download. Model availability, download size, memory, and free-storage requirements vary by device. Chatter checks the exact path on the device rather than promising support from a rough phone-generation list.

If your device supports Apple Intelligence, open Settings › Apple Intelligence & Siri and make sure it is enabled and has finished setup. Otherwise, Chatter may offer a compatible local model from Settings › AI Engine after a verified purchase or restore. That optional model downloads directly from Hugging Face into Chatter's app-local container, not a shared App Group. Keep the app on a reliable connection for the initial download. Once the selected model is ready, conversation and coaching work offline.

Chatter downloads optional conversation-model and natural-voice files directly from Hugging Face. Neither request sends microphone audio, transcripts, conversations, partner replies, or other practice content. Hugging Face may collect data it classifies as CoarseLocation, OtherDiagnosticData, and ProductInteraction, with linked=true and tracking=false, for AppFunctionality, Analytics, and Other purposes. This is the download provider's boundary for both optional paths. Studio AM does not receive that data and receives no product analytics or tracking from Chatter.

Allow Microphone and Speech Recognition for Chatter in system Settings. Chatter requires on-device recognition and does not retain microphone audio; only the resulting transcript and speaking duration can be saved in the local conversation. You can always type instead.

Yes. Speech recognition, conversation, corrections, recaps, and transcript matching run on the device. Chatter has no Studio AM account, ads, or cloud AI server. Studio AM receives no conversation content, product analytics, or tracking. Saved conversations stay in the local app archive and are not synced through CloudKit. Optional conversation-model and natural-voice downloads go directly to Hugging Face under the separate provider boundary described above; App Store purchase or restore and a link you explicitly open also require a connection.

Pro can download optional natural-voice files directly from Hugging Face. The request sends no microphone audio, transcript, conversation, partner reply, or other practice content. Download and storage size can vary. After download, the voice reads partner replies on-device and does not change the microphone path. You can disable it, remove it, use the system voice instead, or leave partner speech off.

Chatter has no general free AI-practice tier. Monthly and annual subscriptions unlock the same Pro feature level. Apple's purchase sheet shows the products actually available to you, their local prices, eligibility, renewal periods, and terms. Pro covers Daily Rehearsal, starting and continuing conversations, checked wording, coaching, metrics, Say It Again, Practice Again, and voice output. A retired lifetime purchase is no longer sold, but verified earlier owners can restore it.

Your saved archive remains available. You can read and search every saved conversation, export all data as readable Markdown plus lossless versioned JSON, or delete individual conversations. Starting or continuing conversations and creating new AI output require an active monthly or annual subscription or a restored legacy lifetime entitlement. Plans, Restore Purchases, app lock, and appearance settings remain available.

No. It compares the words in the on-device transcript with the improved line. Chatter focuses on conversation, grammar, and natural phrasing; it does not score pronunciation or accents.

No generated suggestion is guaranteed. The visible changed words come from a deterministic comparison between the original and improved lines, but the proposed wording and explanation come from a small on-device model. Treat them as a helpful practice prompt, not a grammar textbook.

Open Settings or the lapsed-access archive and choose Export Markdown + JSON. Export is created only when you request it and opens the system share sheet. Delete a saved conversation from its Review or archive action. Deletion and export do not require an active subscription.