// for consultants, coaches and advisors

Client notes and documents, worked on without leaving your laptop

Client conversations and the documents around them are the two things a consultant cannot casually hand to a third-party service — and they are exactly what the useful AI tools ask you to upload.

The situation

  • A recorded call is worth an hour of writing up, and the write-up is the part that gets skipped.
  • The material is confidential by default: an NDA, a client's commercial plans, or simply a conversation somebody assumed was between the two of you.
  • Transcription services keep the audio. Note-taking assistants keep the transcript. Both are a disclosure you have to be able to explain.

How it works

1. Record or load the audio

Speak into the page, or open a recording you already have. The audio is decoded in the browser.

2. It transcribes on your machine

A speech model runs in the tab. The audio is never sent anywhere, and the transcript exists only in the page until you save it.

3. Turn it into minutes

Decisions, actions and owners are pulled out of the transcript into notes you can paste into whatever you actually use.

4. Ask across your documents

The companion document agent indexes proposals, briefs and reports locally, so you can ask what a document said without opening five of them.

The agents that do this

Open one and it runs in this browser — nothing to install, and no account needed to try it. You can also browse everything else people have published.

What never leaves your computer

  • Audio is decoded and transcribed in the browser tab; no recording is uploaded and none is stored on our servers.
  • The transcript and any notes stay in the page until you copy or save them.
  • The downloaded HTML file works offline, which is the simplest possible answer to "where did the recording go".

Why this is structural, not a policy

There is no server-side AI on AgentOp: the model runs in your browser tab, so your documents have nowhere to be sent. The full account of what our servers can and cannot see is on the trust page.

What it will not do

  • The on-device speech model is a small English model. It handles clear speech well and struggles with heavy accents, crosstalk and poor microphones; it is not a substitute for a professional transcription service on difficult audio.
  • It transcribes what it is given — it is not a meeting bot, does not join calls, and cannot record a call you are not recording yourself.
  • Long recordings take real time to process, because the processing is happening on your machine rather than in somebody's data centre.

Will it run on my machine?

The on-device option needs a recent computer with a graphics card, or an Apple Silicon Mac, and Chrome or Edge — phones and iPads cannot run these models at all. Rather than guess, the browser check measures your actual machine in about a second and tells you which models it can run, before anything downloads. If the answer is no, the same agents work with your own OpenAI or Anthropic key instead — at the cost of the text leaving your machine, which for confidential material is the trade-off this page exists to avoid.

Questions

Does the recording get uploaded?

No. The audio is decoded and transcribed by code running in your browser tab. No recording, transcript or note is sent to AgentOp's servers.

Does it join my calls like a meeting bot?

No, and deliberately so. You record or supply the audio yourself and the agent transcribes it. Nothing dials into a meeting, and there is no always-on listener.

What languages does it handle?

The on-device speech model is English. Other languages are not supported on that path today.

Where do the notes end up?

In the page, until you copy or save them. Nothing is stored on our side, so there is no account to export from and no retention period to ask about.

Related