Questions
How does the beta work?
We are inviting 40 people to use Dalm on their Mac before launch and to tell us what works and what does not. Testers get Dalm before everyone else, a direct line to the people building it, and 2 months of Dalm Pro free when paid plans open. In return, use it in your real work for a month, answer a short survey once a week and join one 20-minute call. It starts when all 40 places are filled, and we will write to you when we send the invite. You need a Mac with Apple Silicon and macOS 15.0 or newer.
Does Dalm join my meeting as a bot?
No. There is no participant to admit and nothing shows up in the attendee list. Dalm works from the audio your Mac already plays and hears, so a video call is captured the same way a phone call on speaker or a meeting in the room is captured. Nobody has to let software into the call, and the other side sees no third party.
Where is my audio processed and where is my data stored?
Audio is transcribed on this Mac by default and deleted once the text is ready; if processing fails, it stays on your Mac. The text is a different matter: transcripts, notes, documents and tags are synced to our servers, and when a document or an answer about a meeting is assembled, that text goes to third-party AI model providers.
Does Dalm work without a network?
Recognition does. Dictation and transcription keep running on this Mac with the network off. Two things need the connection and neither is optional: signing in, without which the app does not open at all, and the automatic sync of the text to our servers, which has no off switch. Documents and answers about a meeting are assembled on the server as well, so those wait until you are back online.
How much does Dalm cost and what is free?
There is a free plan. Pro is $16 per month, or $12 per month billed annually. No plan puts a limit on how long anything is kept. Once you subscribe, your price and limits won't get worse for as long as your subscription keeps renewing.
What do I need to run it?
A Mac with Apple Silicon, running macOS 15.0 or newer. On first launch the app downloads a speech model before it can transcribe anything, so the first run takes longer than every run after it, and the same happens once more when the first conversation is processed. Exact sizes are named in the onboarding, before anything starts downloading.