Your voice stays on your Mac.

What Speechcap does NOT collect

• Audio recordings.On Pro with the on-device engine, your audio is captured in memory, transcribed locally, and discarded. It is never written to disk, never uploaded, never reaches our servers. On the cloud engine (Free tier and Pro's optional cloud mode), audio is streamed transiently to our transcription provider, processed, and discarded — we do not retain audio.
• Transcripts in our systems. The text Speechcap produces is written into the focused app on your Mac (Slack, Mail, Cursor, anywhere) and a local history database on your machine. We do not upload transcripts to our servers.
• Behavioural tracking across apps.Speechcap doesn't watch what other apps you run, what you copy and paste, or what you read on the web.
• Browser/page analytics fingerprints. The marketing site (speechcap.com) uses privacy-respecting analytics only — no third-party advertising trackers, no cross-site behavioural fingerprinting.

What Speechcap DOES collect

• Account information. If you create a Pro account, we store your email address, a hashed password, your subscription state (active, trialling, cancelled), and the country you signed up from (for PPP-localised pricing).
• Word-count usage on Free.To enforce the 2,000-words/month Free quota, we count transcribed words per user per calendar month. The counter is a single integer — we don't store the words themselves.
• Crash diagnostics.If Speechcap crashes, we may receive an anonymised crash log (no audio, no transcripts, no personal data) so we can fix the bug. You can disable this in the app's Settings → Privacy.
• Custom vocabulary you've opted to sync.If you turn on vocabulary sync, the list of custom words you've added (names, acronyms, product codes) is stored against your account so it follows you between Macs. You can turn sync off; the list then stays local-only.

Cloud transcription mode (Free + optional on Pro)

On the Free tier and on Pro's optional cloud mode, audio is transmitted to a transcription service (Cloudflare Workers AI / Groq, depending on availability) over TLS. The audio is processed and then discarded by the transcription provider. We do not retain audio at any stage; the providers' own policies govern any internal retention windows for abuse-prevention purposes.

If you don't want any third-party processing of your audio: upgrade to Pro and select the on-device engine in Settings → Transcription. That setting routes all transcription and AI cleanup through models running locally on your Mac. Nothing leaves the machine.

Subprocessors

We use a small number of third-party services to operate Speechcap. None of them receive your audio or your transcripts. They process metadata only.

• Supabase — authentication, account database, billing webhook handler.
• Polar — subscription billing and invoicing.
• Cloudflare Workers AI / Groq— cloud transcription (only when you're using cloud mode, and audio is transient).

Your rights

You can:

• Export your account data (custom vocabulary, account email, subscription history).
• Delete your account (this removes your account, custom vocabulary, and stops all future cloud calls).
• Disable cloud transcription entirely by switching to the on-device engine on Pro.
• Disable crash diagnostics in the app's Settings → Privacy.

To exercise any of these rights, email hello@speechcap.com. Requests are usually fulfilled within 7 days.

Changes

If we change anything material about how we handle your data, we'll update this page with a new “Last reviewed” date at the top and announce the change in the app and via email to active Pro subscribers.

Contact

Privacy questions, data requests, or anything else: hello@speechcap.com.