No extensions. No desktop app. No audio routing. No IT approval. Pertch runs entirely in your browser: open a tab, share your meeting audio, and start getting coached.
Navigate to app.pertch.app in any Chromium-based browser. Sign in with the 6-digit code we email you. No password to remember, no account setup friction. You'll be in your coaching dashboard in under 10 seconds.
Before your call starts, select one of three coaching lenses. You pick the one that fits the call ahead of you.
Each methodology tunes the AI to look for different patterns in the conversation. Decisive watches for hesitation and missing next steps. Qualify looks for gaps in pain, decision-maker, and timeline. Screen helps you qualify fast on high-volume calls: budget, authority, need, and timeline. (Reframe and Consultative are in development.)
Click "Start Coaching" and Chrome will prompt you to share audio. Share the browser tab your call is in, or share a window or your whole screen with system audio if your meeting runs in a desktop app like Zoom. Pertch also captures your microphone so it can tell your voice apart from the prospect's. No virtual audio devices, no desktop install.
As the conversation unfolds, Pertch keeps a running memory of the whole call and weighs the latest exchange against it. When it detects a coaching-worthy moment (hesitation, an objection, a missed qualification question), a short, contextual prompt appears in your coaching HUD.
When your call is done, click "End Session." Your audio is gone for good. It was never recorded. Your transcript, a post-call summary (objections, next steps, key takeaways), and every prompt you saw are saved to your private session history for 30 days, so you can review and learn from them. After 30 days, the raw content is permanently deleted.
Your shared audio (a tab, a window, or your whole screen) and your microphone stream directly to Pertch. No recording, no disk writes, no virtual audio devices. The streams are ephemeral. They exist only in memory.
Audio chunks are transcribed using a streaming speech-to-text model optimized for conversational English. Latency: under 2 seconds from spoken word to text.
The latest exchange is evaluated against a running memory of the entire call and your selected methodology. The model looks for specific patterns: missed questions, objection signals, hesitation markers, qualification gaps.
If a coaching-worthy moment is detected, a single tactical prompt is generated. It must be one sentence, actionable, and specific to what was just said. If there's nothing useful to say, Pertch stays silent.
The prompt appears in your coaching overlay. You glance at it, decide whether to use it, and optionally give thumbs up/down feedback. Total time from spoken word to visible prompt: 5-15 seconds.
Any Chromium-based browser, version 100 or later. Tab audio sharing requires the Chrome Tab Capture API.
A standard broadband connection. Pertch streams audio to the transcription engine in real time, similar bandwidth to a video call.
Zoom, Google Meet, Microsoft Teams, Webex, dialers. Share the browser tab if your call runs in one, or share your whole screen with system audio to use a desktop meeting app.
No extensions, no plugins, no desktop app, no IT approval, no admin permissions, no virtual audio drivers.
Start coaching in under 2 minutes. No credit card required.
