Skip to content
Platform Documentation

Connect AI

Connect AI adds speech and text assistance to messaging and related workflows. Availability depends on your Thirdlane Connect version and tenant configuration. Feature toggles live under Settings → Connect AI.

Connect AI settings page

Open Settings, then Connect AI. Typical controls include:

  • Feature toggles — Turn capabilities on or off according to policy.
  • Engine status — See which AI or speech engines are active or need setup.
  • Model size / engine selection — Choose an appropriate balance of quality and resource use where offered.
  • Microphone test — Verify input before dictation or voice messages.
  • Dictation test — Confirm speech-to-text behavior.

If Connect AI does not appear, your tenant or branding policy may hide it, or your role may not include access.

Voice messages

In direct conversations and channels (where enabled):

  1. Use the microphone control in the composer to record a voice message.
  2. Preview the recording and optional waveform.
  3. Adjust playback speed (for example 1x, 1.5x, 2x) when playing messages you receive.
  4. When speech recognition is enabled, messages may include an inline transcript for search and skimming.

Mobile: Recording and playback vary by platform; web and desktop have the fullest support described here.

Live dictation

Dictation streams your speech into the message draft:

  • Uses the browser’s speech recognition where available, or an on-device engine when configured.
  • Desktop (Electron): On-device recognition is often required because the browser speech API may not run inside the desktop shell—install optional on-device speech assets when prompted.

Stop dictation before sending if you need to edit the text.

Rewrite with AI

Some builds include a rewrite or sparkles control near the composer:

  • Offers options such as Improve, Tone, or Length.
  • Adjusts draft text without auto-sending—you review before you post.

Exact labels depend on your version and tenant settings.

Transcripts for voice messages

For incoming voice messages:

  • On-demand transcript — Generate text locally when supported, with copy-to-clipboard.
  • On-device Whisper-style recognition — Optional download for offline-capable transcription in supported languages.