Complete one small, verifiable task before using the workflow for a larger project.
Use the current product from the official source. Check its device requirements and entitlements; a free download does not necessarily mean every operation is free. The feature table records product-specific limits.
Check release availability on the official site before planning a trip around the app. The site currently identifies it as an upcoming App Store release.
When a supported build is available, confirm the device/OS requirements and activate the required subscription while online.
Choose both conversation languages. Select the mode deliberately rather than assuming every mode has the same connectivity and data handling.
If you intend to use On Device, download the compatible models and confirm offline speech recognition for both languages in advance.
Allow microphone access and test a short exchange in a quiet room. Check original text, translated text and spoken playback.
Take turns with the other speaker. Use short sentences for names, numbers and instructions; correct the original transcript if speech was misunderstood.
Stop the session when finished. Review the history and credit balance. Re-test the planned offline workflow with connectivity unavailable before relying on it.
Return to the saved output or record rather than relying only on the preview. Check the specific items described in the steps above. If the result differs, keep the original input and reproduce the issue with a small, non-sensitive example.
Choose a scenario from practical workflows, or read features and limitations before expanding the task.
Documentation home · Official product website
Reviewed 2026-11-04 against public product materials. Interface wording and availability can change with the installed version.