Your whole source archive, never in someone else's cloud.
Transcribe interviews and source tapes on your own machine, search years of recordings, and ask the archive questions with quotes you can play back — then wipe a source yourself when you need to. No upload means no vendor's retention policy between you and your sources.
Every interview transcribed with speakers, filed in a private library on your machine.
Authors, biographers, and documentary researchers with years of interview tape run the same archive loop: one owner, whole-archive questions, and a wipe that takes the chats with it.
The jobs
Find the receipt across years of tape
Ask a cross-file question — what did this source say about the contract in 2022 versus 2025, what did the agency say about the subcontractor — and get verified quotes with a jump to the moment. “2022” means when the tape was recorded: Kajo reads the date from the file — its embedded creation date, or else the file's own date — and you can correct it.
Check the story against the archive
Ask whether anyone else described the same warehouse, where sources agree, and where they diverge. Evidence first: citations are how you check yourself before you publish — and a plain “no supporting evidence” is how you learn the archive does not say it.
Translate the interview, then ask in English
Import a foreign-language tape, transcribe it, translate it on-device, and choose Ask this recording in English. Quotes stay in the original language with the original timestamps, and bilingual SRT or VTT subtitles export when the piece needs them.
Brief yourself before re-approaching a source
Ask for a brief on that source's folder: what they have already said, what changed between interviews, and what is still unasked — every point cited, so you walk in knowing your own tape.
Delete a source yourself
After publication, or when a source is burned, deleting the recording removes the app's audio copy, transcript, search index, and every chat pinned only to it, and strips its citations from other chats — delete those too if their answers quote the source. It works offline and there is no vendor retention ticket to file. Kajo does not overwrite freed disk space, so keep the disk encrypted and mind your backups.
Source protection is a custody problem
Cloud transcription turns source audio into someone else’s retention policy. Even when a vendor promises deletion, the recording has already crossed a network and sat on infrastructure you do not control — reachable by their staff, their subprocessors, and whatever legal process reaches them.
Kajo keeps the custody chain short by construction: transcription runs on your machine, transcripts sit in a local database on your disk, and cited chat never uploads transcript text. That is the difference between a privacy policy and an architecture.
Kajo vs cloud transcription for journalists
Feature
Kajo
TurboScribe
Rev
Where the audio is processed
On your machine
TurboScribe’s cloud
Rev’s cloud
Where the archive lives
Your machine
Your TurboScribe account (cloud)
Your Rev account (cloud)
Asking the archive, with citations
On your machine
Chat with all transcripts inside ChatGPT via "TurboScribe GPT", free; no chat inside TurboScribe itself
Cited answers across all your files, in Rev's cloud
Wiping a source
Delete it in Kajo; the original file stays on your disk
Files are kept while your account is active unless you delete them; a deleted file or account is fully purged, backups included, within 90 days
Delete files from your Rev.com account at any time
Works with the network off
Yes
No
No
Linux
Native app
Web app only — no desktop or mobile apps
Browser only
Price
$49 launch / $99 standard, one-time
Monthly or yearly plan
Per minute or per seat
A reporting workflow that stays local
Download Kajo for Mac, Windows, or Linux — Free needs no account.
Import interview files, or on Lifetime point a watch folder at your recorder’s export directory.
Let them transcribe, check names and speaker labels, and file each story or source in its own folder.
Search or ask across the archive when you need a quote, a timeline, or a cross-source check.
After publication, delete what needs protecting — no vendor retention ticket.
Free
5 files or 150 minutes of audio — total, for life. Core loop (transcribe, translate, summarize, chat), no card, no upload.
✓Transcribes 99 languages on your laptop — all on-device.
No. Transcription, speaker labels, translation, and chat all run on your machine — Kajo never uploads your audio or transcript text.
Can I ask my whole archive, not just one folder?+
Yes. A journalist's archive has one owner, so searching the whole library is a normal question: choose Search entire library… in Chat and confirm. Story or source folders still let you narrow the question when you want to.
How does this compare to TurboScribe or Rev?+
Both are cloud services: the audio is uploaded and kept in your account on their servers, and chat happens in their cloud or, for TurboScribe, through ChatGPT. Kajo transcribes on your machine and keeps every transcript in a private library you can search and ask with citations.
Can I wipe a sensitive source after publication?+
Yes. Deleting a recording removes Kajo's audio copy, its transcript and search index, and chats pinned only to that recording; other chats lose their citations to it but keep their answer text. The original file on disk is not deleted — remove it yourself.
Does Kajo work offline?+
Yes, once the models are downloaded: transcription, translation, and cited chat all run with the network off, which suits field work on unreliable connections and machines you keep offline.