Skip to content

Voice intake (Hold to Speak)

Hold to Speak Lead turns a short voice memo into the same structured draft you get from Magic Paste. It is included on Foreman, Autopilot, and Founding 50 — not on Climber or sandbox trial.

  1. Open Dispatch → Lead intake (Magic Paste).
  2. Press and hold Hold to Speak Lead while you describe the job (customer, address, scope, equipment).
  3. Release when done — ArborBuddy transcribes the clip, then runs the normal AI parse.
  4. Review the draft in the resolve/review flow before approving to the board.

A waveform animates while you hold the button. Transcription usually takes a few seconds after release.

  • Speak clearly; 15–60 seconds is the sweet spot.
  • Include address, tree species, and scope (“big oak in back, tight yard, need crane”).
  • Wind and engine noise are common — re-record if the transcript looks wrong; you can edit text before parsing.
  • Each successful voice parse counts as one lead intake (same monthly quota as Magic Paste — see Plans & tiers).
  • Clips are capped at 90 seconds and 5 MB.
  • Mic permission is required in your browser.

Audio is sent to Cloudflare Workers AI for transcription only. The app stores the text transcript in the intake audit trail — not the raw audio file in v1.

IssueWhat to do
Button missingUpgrade to Foreman+ in Settings → Billing.
“Microphone permission denied”Allow mic access for app.arborbuddy.com in browser settings.
“No speech detected”Hold longer, speak closer to the mic, or paste text instead.
Quota exceededWait until next month or upgrade for a higher AI import limit.

Approved voice leads follow the same path as pasted leads — see Review queue if the draft needs a second look.