For Institutions

AuTrans for Production Companies and Content Studios

Captions, subtitles, show notes and a searchable archive for teams publishing every week — including dialogue in Pidgin, Yoruba, Hausa and Igbo.

A head of production or a content operations lead, spending against a client project or a channel budget rather than an annual software line. They decide quickly, but only after running one of their own difficult episodes through it — a noisy set, a three-language conversation, a guest nobody else's tool has ever handled.

What this looks like today

  • Captioning a week of episodes by hand takes an editor off the edit, and the publishing calendar does not move
  • Cutting long-form into clips means scrubbing the whole recording to find the thirty seconds worth cutting
  • Dialogue in Pidgin, Yoruba, Hausa or Igbo gets captioned by whoever on the team speaks it, when they have time
  • Platform auto-captions mangle Nigerian names and code-switching, so they either ship wrong or get switched off
  • The back catalogue is unsearchable, so a line a guest said two seasons ago is effectively lost rather than reusable
  • Outsourced captioning is quoted per minute, so cost rises with output while the margin on a channel does not

What changes

  • Episodes come back as SRT and VTT alongside the text, so a caption pass becomes correction rather than typing
  • Every line carries its timestamp, so a clip-worthy moment is found by searching the transcript and cut from the timecode
  • Pidgin, Yoruba, Hausa and Igbo are transcribed directly, and a line that switches language mid-sentence stays as it was spoken
  • Speaker labels keep a multi-host show or a multi-character scene readable, and renaming a speaker once applies throughout
  • AI notes turn a finished episode into show notes, topics and pull quotes while the edit is still open
  • Translation flattens a mixed-language episode into one language for a subtitle track aimed at a different market
  • Search across the whole archive, so old episodes become source material instead of files nobody can open

Procurement, data and contracts

  • Invoiced per project or monthly, in Naira, against a purchase order — no per-seat licence to reconcile for a team whose size changes with the slate
  • Client audio is never used to train AI models. Text corrections are only ever used where an account has explicitly opted in, and a studio account can have that off entirely
  • Everything for a project can be deleted on request, including the derived data — or held for a fixed period if a client contract requires retention. Tell us which, and the window
  • View-only links let a client, a broadcaster or a platform reviewer read a transcript or check a caption pass without an account and without occupying a seat
  • There is no developer API yet — it is on the waitlist at /developers, not shipped. Delivery today is dashboard upload, several files at a time, and SRT, VTT, DOCX, PDF or text out
  • Allowances are set on the account directly, so a heavy delivery month does not need a plan change and a quiet one is not paid for twice
  • We are CAC-registered in Nigeria and invoice in Naira, which keeps a production budget clear of forex cards and foreign-vendor onboarding

Frequently asked questions

Our episodes mix English and Pidgin, sometimes Yoruba. Does that break it?

No — that is the case AuTrans was built for rather than an edge case. A conversation that moves between Nigerian English, Pidgin, Yoruba, Hausa or Igbo is handled as one recording, in the languages actually spoken, so nobody has to split a file by language before uploading it.

Do we get subtitle files, or just a transcript?

Both. Every recording exports to SRT and VTT for captioning and platform delivery, and to TXT, DOCX or PDF for scripts, show notes and archives. The transcript stays editable with the audio playing alongside it, so an editor fixes a name or a timing and exports the corrected file.

How fast is it across a week of episodes?

Minutes rather than hours per file, though the exact time depends on the recording and on load at that moment. The practical change is that captioning stops being the step that decides whether an episode ships on schedule, and several files can be queued at once instead of moving through one editor.

Our clients ask what happens to their footage.

Audio is never used to train AI models, under any plan, including institutional ones. Text corrections can be used to improve the product, but only where an account has explicitly opted in — and a studio account can simply have it off. Recordings stay private to your account unless you deliberately share a view-only link, and everything for a project can be deleted on request including the derived data.

Can we pay per project instead of committing to a year?

Yes, and for most studios that is the right shape. Per-file pricing starts from $0.20 with no subscription, and we also invoice monthly or per project against a purchase order. If your volume turns out steady enough that an annual arrangement costs you less, we will say so rather than let you overpay per file.

Can we test it before putting it in a delivery pipeline?

Yes. The free plan includes 30 minutes of transcription every month, which refills — enough to put your hardest audio through it before anyone signs anything: the noisy set, the three-language conversation, the guest with the accent that defeats everything else. That is a stretch of an episode rather than a full slate, so if you want to run a whole show or a delivery week end to end, ask and we will set the allowance for it. Judge it on your own material rather than on a sample we picked.

Related

Tell us your volume and we'll quote

Roughly how many hours of audio a year, and in which languages. We price against that rather than a published tier. You can also start on the free plan first and run your own recordings through it before anyone raises a purchase order.

Talk to us