Features
Automatic AI Responses in Meetings
This is Pluely's signature feature. While you're in a live meeting, interview, or client call, Pluely listens to the conversation and responds automatically — a suggested answer when the interviewer asks you something, a live translation as the other person speaks, running meeting notes as the discussion unfolds. You decide when it responds; your selected prompt decides what it says.
No bot joins the call. No one sees anything but you.
What powers each response
Every automatic response is generated from three things, combined into a single AI request:
- Your selected prompt — the system prompt you picked (Interview Assistant, Meeting Assistant, Translator, Technical Interview, or any custom prompt you wrote). This fully defines the kind of response you get.
- Your prompt's attached files — documents you attached to that prompt (your résumé, a product spec, pricing sheets) ride along as reference knowledge, so answers are grounded in your material. See Prompt presets.
- The live transcript — the most recent lines of the actual conversation, speaker-labeled.
So with the Interview Assistant prompt selected, a question from the interviewer produces a natural, speakable first-person answer. With the Translator prompt, every pause produces a translation. With a custom "sales objections" prompt and your pricing PDF attached, an objection on the call produces a grounded rebuttal. Same trigger machinery — completely different copilots.
When Pluely responds: the Auto responses control
In Listen mode you'll see a chip labeled Auto responses · On questions (or Every pause / Off). Tap it to configure the trigger:
- When a question is asked (default) — Pluely detects questions in the conversation and responds to each one. Detection is smart: it catches questions even when the transcription drops the question mark ("tell me about your experience with…", "walk me through…", "have you considered…"), and it runs entirely on your device — detection itself never costs anything.
- After every pause — the continuous copilot. Whenever a speaker finishes and the conversation goes quiet for a moment, Pluely responds to what was just said. Ideal for live translation, running notes, and accessibility captions-with-context. This mode generates more responses, so it uses more of your monthly AI requests — the control tells you exactly how many you have left.
- Off — nothing fires automatically. The Suggest button and quick actions still work on demand.
Your choice is remembered across sessions and syncs between the overlay and the dashboard.
Pace: how quickly it jumps in
The same popover sets the pace:
| Pace | Waits after speech | Minimum gap between responses |
|---|---|---|
| Fast | ~1 second | ~5 seconds |
| Balanced (default) | ~1.5 seconds | ~15 seconds |
| Relaxed | ~2.5 seconds | ~40 seconds |
The wait is a true quiet window — Pluely watches the live transcription and only responds once the speaker actually stops, so it never answers mid-sentence. If someone asks a question and then keeps talking, Pluely holds until they finish and answers the complete thought.
Built-in cost protection
Automatic responses are designed to never waste your monthly AI requests:
- Short remarks never trigger a response. "Yeah." "Okay, got it." "Sounds good." — backchannels and small talk accumulate silently; a continuous-mode response only fires once enough new conversation has actually happened.
- Repeated questions aren't billed twice. If the same question echoes through the transcription again moments later, Pluely recognizes it and skips it.
- One response never overlaps another. If a new trigger lands while an answer is still streaming, Pluely queues it and responds once — never in parallel.
- Detection is free. Question detection and pause detection run locally on your device. You only spend an AI request when a response is actually generated.
- You always see the budget. The Auto responses popover shows exactly how many AI requests you have left this month, live.
Reading the signals
- The Auto responses chip pulses while a response is queued — the copilot is waiting for the speaker to finish.
- Each response appears in the answer panel with the question it's responding to shown in the header, so you always know what it's reacting to.
- Responses join the session's conversation history, so a manual follow-up ("make that shorter") builds on what was already said.
Manual control is always there
Whatever mode you pick:
- Suggest generates a response from the recent conversation on demand.
- Quick actions / follow-up chips send a specific instruction with the transcript as context.
- Typing your own instruction works at any time.
Set it up in 30 seconds
- Open Listen mode and pick your prompt (Interview, Meeting, Translator, or your own) from the prompt picker.
- Tap Auto responses and choose when Pluely should respond — and how eagerly, with pace.
- Press Start listening. That's it — answers appear as the conversation needs them.
Related
Last updated 2026-07-10 · pluely.com