Help that arrives before the conversation moves on
Every meeting tool is useful tomorrow. Almost none are useful in the four seconds after a question lands. Whisperer builds an answer from the running transcript, your knowledge base and your profile — while the other side is still finishing the sentence.
Nothing joins the call · nobody else sees the window
The answers below are an example. Your own base connects in the app.
…
…
Let's start with Q3 — how did conversion move?
Conversion is up 14% — let me pull up the funnel.
And the cost per lead?
Cost per lead is €12.40 — 9% under plan. Most of the gain came from organic.
Not a video — this is Whisperer itself, and the panel really answers. The microphone starts only when you press it.
Why not a chat
An isolated question gets an isolated answer
You could paste the question into a chat window. What comes back is a good general answer to a general question, because that is all the chat was given. It does not know what was said two minutes earlier, what you charge, or which of your projects is the relevant one.
The difference is not the model. It is what the model was handed.
- You retype the question, losing how it was actually asked
- No idea what the previous ten minutes established
- No access to your numbers, projects or past calls
- You look away from the call to read it
- Takes the question from the transcript, as it was said
- Carries the whole running call as context
- Mixes in your notes and your profile before answering
- Appears in a window above the call, not in another app
Modes
Different questions need different answers
Limits
Two things worth knowing before you rely on it
A finished recording can be uploaded: the Meetings section takes one file up to 25 MB — mp3, m4a, wav, ogg, flac, aac, plus mp4 and webm, from which the audio track is used. What is missing is bulk upload of a whole archive; and it spends the same minutes a live call does.
The advantage over a chat window comes entirely from your material. With nothing in the knowledge base, you get what a chat window would have given you — this is the most common reason people are underwhelmed in week one.
Questions
What people ask first
How do I ask without typing?
Press enter on an empty field and Whisperer takes the last thing the other side said as the question. That is the path built for the case where the question was just asked out loud and you have no hands free.
Is the answer read out to me?
It appears as text in the window, streaming as it is generated, so you can start reading before it finishes. Reading it aloud is not the intended use and is audible to the other side.
Does the other side notice anything?
Nothing joins the call and the window is excluded from screen capture. What they may notice is you reading — which is a reason to use the answer as a footing rather than a script.
Which model answers?
The one set for that role, from a catalogue spanning OpenAI, Anthropic, Google, DeepSeek, Qwen, Moonshot, xAI and NVIDIA. Frontier models are part of the Max plan.
Is this allowed?
It depends entirely on the conversation. In sales, negotiation or a customer call, using your own material is unremarkable. In an assessment of a skill, the line is different and worth reading about before you decide.
The only test that matters is a real call
Put five notes about your own work in the knowledge base first. Then take a call you would have taken anyway, and see whether what appears is worth looking at.