Call Transcription

How transcription works, how to enable it, and how to review transcripts alongside your notes.

The hotline can automatically transcribe phone calls into text. What makes this different from most transcription services is that the audio never leaves your device. Transcription happens entirely in your browser.

How it works

When a call ends, the audio is processed by a speech recognition model running directly on your computer. The resulting text is then encrypted and saved alongside your call notes. At no point does the audio travel to an external server.

This design protects caller privacy — even the hotline server never hears the audio.

For operators: enabling transcription

Go to Settings and find the Transcription section. Toggle the global switch to enable transcription for your organization.

Even when enabled globally, individual staff members can opt out through their own preferences (see below).

For staff: your transcription preferences

Go to Preferences to control whether your calls are transcribed. If you turn transcription off, your calls will not be transcribed regardless of the global setting.

You might want to disable transcription if:

• Your device is older and transcription slows it down
• You prefer to write notes manually
• A caller has asked not to be recorded

Reviewing transcripts

Transcripts appear alongside call notes. When you open a call from the Calls page or the Notes page, the transcript (if available) is displayed alongside your notes for that call.

This makes it easy to cross-reference what was said with what you wrote down.

Language support

The transcription model supports multiple languages. It will attempt to detect the language being spoken automatically. The quality of transcription varies by language — common languages like English and Spanish tend to produce better results.

Privacy note

Because transcription happens on your device, it requires some processing power. Older computers or phones may experience slower performance during transcription. The audio data is discarded after the transcript is generated — it is not stored anywhere.