Now in open beta

Hear everything.
Say the right thing.

conva listens to both sides of your live call, transcribes as it goes, and puts the right answer on screen — before the pause gets awkward.

Join the beta See how it works

Audio never leaves your machine · Windows & macOS · Free while in beta

An instrument, not a notetaker

Everything happening at once, while you talk.

Dual-channel capture

Their audio and yours on separate channels — your mic and the system output — so nothing is attributed to the wrong voice.

Live transcription

Speaker-labelled text in under a second, streamed into a dual-column chat, timestamped and searchable the moment you hang up.

AI assist, mid-sentence

Suggested replies and summaries arrive while they’re still talking — not in an email afterwards. One click to research any line.

Grounded in your docs

Answers are retrieved from your own reference library (hybrid RAG) and cited back to the source — never a confident guess.

Saved conversations & recording

Keep any call as a named conversation and record it to a clean stereo WAV — you on the left, them on the right.

Private by default

Speech-to-text runs locally, on your device. Nothing is stored on our servers unless you save it. Cloud engines are strictly opt-in.

How it works

Both sides in. The right answer out.

THEM
YOU
Capture

It hears both of you

conva taps your microphone and the system audio at once, keeping each voice on its own channel.

00:14So what happens if we onboard a second team mid-contract?
00:21Good question — let me pull up how seats work.
00:26And is that prorated, or…
Transcribe

Live, dual-column text

Every line is transcribed in under a second and labelled by speaker, so you can read the room as it happens.

“Honestly, the price is higher than we budgeted.”
AI
SAY THIS
Reframe on cost-per-seat, then offer the annual plan — it lands 18% under their budget.
Pricing Playbook · p.4
Assist

A grounded answer, in the moment

The AI drafts a reply from your own documents and cites where it came from — you decide whether to use it.

Private by default

Your calls stay on your machine.

conva was built local-first. Capture and speech-to-text run on-device with an offline model — no upload, no account required to listen.

  • On-device transcription — audio is processed locally and never leaves your computer.
  • Nothing is stored on our servers unless you explicitly save a conversation.
  • Cloud speech and LLM engines are opt-in — your keys, your provider, your choice.
Powerful, and flawless on mobile

Desktop today. In your pocket next.

conva is one product across desktop and mobile, built from a shared core. The desktop app ships now for Windows & macOS; a mobile companion is on the way — designed to feel powerful and read perfectly on phones of every size.

Mobile companion — coming soon

Pricing

Simple while it’s early.

conva is free while it’s in beta. When it launches it’s one plan — and beta users lock in the launch price for good.

Free while in beta

Pro includes a generous monthly assist meter. Need more in a heavy month? Top up with credits — no plan change, no surprise upgrade.

Pro LOCKED IN FOR BETA USERS
$29/ month

Unlimited live calls for one seat, with a monthly assist meter.

  • Real-time assist, grounded in your documents
  • On-device transcription & call recording
  • Searchable, saved conversation history
  • Monthly assist meter + credits to extend anytime
Join the beta — lock in $29

Your next call goes better.

Free while in beta. Windows & macOS. Nothing to configure — open it, join your call, start talking. Drop your email and we’ll send your invite.

No spam — just your beta invite and the occasional product note. Privacy.