About Textspert
Textspert turns recordings into text you can edit, search and share. We built it because we wanted transcription that didn’t mean handing our recordings to an AI company.
Why we built it
Recordings can be sensitive: an interview, a client call or a voicemail from family. We wanted a tool where the answer to “where does my audio go?” is short.
We run open speech and language models on GPU servers we operate. The private transcription page explains our training, access and deletion promises.
How it’s built
Speech recognition uses Parakeet for 25 European languages and Whisper large-v3-turbo for everything else. Sortformer labels speakers. An open Qwen model writes summaries, answers chat questions and handles translation alongside MADLAD-400. We picked these after testing them side by side on the same recordings, and we keep the settings that did best, not the ones that sound impressive.
What we publish and what we don’t
- Every limit: 60 minutes per Free file, 4 speakers per file, 3-hour speaker blocks, and fair use up to 720 hours a month on Unlimited.
- What other services do better, on our comparison pages.
- No accuracy percentages, because we haven’t published a test that would back one up.
- No customer logos, ratings or user counts. When we have real ones, we’ll show them.
What Textspert doesn’t do (yet)
We don’t join live calls, so record the meeting and upload it yourself. You can’t import a YouTube or Dropbox link; upload the file. There’s no human proofreading or certified transcript. You can check and edit the machine transcript in the browser. Captions come as SRT and VTT files. They aren’t burned into your video. There’s no developer API yet.
Some of these are on the list. If one of them is the reason you can’t use Textspert, tell us; it helps us decide what to build next.
Talking to us
Questions, bug reports and feature requests all go to the same inbox. The contact page has the address and what to include so we can help on the first reply.
About the company
Is Textspert free to try?
Yes. Free covers 3 files a day, each up to 60 minutes, with no card.
Who operates the servers?
We operate the GPU machines that run transcription and the AI tools. No outside AI service sits between your file and the transcript.
Do you offer an API?
Not yet. Textspert is a web app for now.
Where to go next
- Private transcriptionYour files, sharing and deletion stay in your control.
- PricingFree, subscriptions, and one-time hour packs.
- CompareTextspert next to TurboScribe, Otter, Rev, Descript and more, with sources.
- ContactEmail support for billing, files and privacy.
Compare plans on the pricing page, or read our short privacy promises.
Join the waitlist
Launching November 2026. Be first to hear when Textspert opens.
We'll only email you about the launch. Unsubscribe anytime. Privacy