An AI workspace, not five apps that never talk

A transcription tool, a notes app, a task board, a knowledge base and a chat assistant. Five subscriptions, and none of them knows what the others know. Whisperer is the same five things sharing one context.

Open a workspace Start with meetings

macOS, Windows and the web · 60 free minutes a month

One contextWhat each part already knows

Meeting
Heard the deadline being moved
03 Apr
Board
Has the task that came out of it
same hub
Knowledge
Holds the contract clause it depends on
note
Assistant
Answers using all three at once
live

The problem

Context does not survive the gap between tools

Every integration between two products moves data. None of them moves understanding. A summary pasted into a task tracker is a summary in a task tracker: the assistant that reads it later has no idea which call it came from, what was decided just before, or which document the decision depended on.

That gap is why the same question gets answered from scratch every time — and why the answer keeps being generic.

What lives in the workspaceeach one is a page

Conversations Meetings Live transcript, real-time answers, and a map of what was decided. What you know Knowledge base Notes, wiki-links and a graph that get mixed into answers. What you owe Tasks A board per piece of work, fed by decisions from the calls. Who acts Leo, the agent Reads all of it and can create, move and schedule on request. How you get back Search Plain-language questions across everything you have said and saved. From outside MCP Claude, ChatGPT and Cursor reading the same context, with consent.

Models

One subscription, a model chosen per job

Answering in a live call, writing code, reading a diagram and drawing a system design are different jobs, and the best model for each is rarely the same one. Whisperer keeps a model per role rather than one model for everything, and you set them.

OpenAI
Anthropic
Google
DeepSeek
Qwen
Moonshot
xAI
NVIDIA

Frontier models are included on the Max plan; the other plans use the rest of the catalogue. You can also write the prompt for a role yourself, attach skills to it and choose what context it is allowed to use.

Limits

What a workspace still cannot do

Limit one It starts empty, and empty is generic

Shared context is only worth something once there is some. The first week of answers will be unremarkable, and there is no honest way around that except using it.

Limit two One recording at a time, not an archive migration

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.

Questions

What people ask first

Do I have to move my whole workflow into it?

No, and most people should not start that way. The meeting side works on its own; the board, the knowledge base and the agent become useful as material accumulates. Notion, Slack, HubSpot and Zapier connect if you want to keep your existing destinations.

Is it a desktop app or a web app?

Both, doing different halves. The desktop client for macOS and Windows handles the live call, because that needs system audio and a window above the screen. Everything after the call lives in the web account.

Who can see what is in my workspace?

You, and whatever you explicitly share. External AI clients reach only the meetings you marked as shared, only through MCP, and only in the areas you approved.

What does the free plan include?

60 minutes of recognition a month, so you can see the mechanism work. The knowledge base and the advanced answer modes open on Start, Pro and Max, which also remove the minute limit.

Start where the pain is

Nobody adopts a workspace on day one. Run it on the meetings you already have, and let the rest earn its place as the context fills in.

Start free