Skip to content
Browse the docs

Realtime and Post-Process+ Recording Modes

Choose live transcription or higher-accuracy processing after recording, and understand what can be changed while a session runs.

Kalima offers two ways to turn a new recording into text. Realtime shows the transcript while people speak. Post-Process+ records first and produces the transcript after you stop. Both use the same amount of transcription time.

Compare the modes

RealtimePost-Process+
Transcript timingAppears while people speakUsually arrives a few minutes after stopping
AccuracyDesigned for immediate, reliable live textHighest accuracy and best speaker separation
Live translationAvailable during recordingApplied during processing
Live viewersCan follow the transcript as it happensSee the transcript after processing
Languages and contextSet before recordingCan be adjusted until you stop

Choose Realtime for captions, live translation, or viewers who need to follow along. Choose Post-Process+ when the final transcript matters more than seeing text during the conversation.

While Post-Process+ is recording

No transcript appears during the conversation. Kalima keeps recording the audio, and you can still adjust languages, translation, names, terms, and session context until you stop.

You may switch a running Post-Process+ recording to Realtime once. Kalima starts live text from that point onward and processes the earlier portion after the recording. You cannot switch back to Post-Process+ during the same take.

After you stop

Kalima uploads and processes the recording in the background. You can close the session while it runs and return later. A notification tells you when the transcript is ready. If processing fails, the recording remains safe and the Studio offers a retry.

Note

Post-Process+ needs a stable upload after recording. Do not remove the local recording data while an upload or retry is still pending.