Short answer: There is no single best transcription tool for journalists — it depends on where your interviews happen and how sensitive your sources are. Good Tape and meaty lead on EU-hosted privacy, Rev on accuracy for rough audio, Trint on newsroom workflow, Otter on live call transcription. One disclosure up front: meaty is our tool, and it records interviews free from any phone browser.
How we compared these tools
An honest note on method: we haven't run these seven tools through a lab test. This comparison is built from each vendor's own published pricing, security, and help pages, plus reported facts from public sources — all checked in July 2026 and linked next to the claim they support. Where a vendor doesn't publish a number, we say so instead of inventing one. And since one of the seven is ours, every entry names a real caveat, including meaty's.
The axes we compared are the ones that matter for reporting, not for corporate meetings:
- Source protection and data handling. Where does your audio physically go — EU or US servers? Is it used to train AI models? Who can hear it, and when is it deleted? For an interview with a vulnerable source, this outranks every feature.
- Speaker labels. A transcript that doesn't say who said what is a liability when you're attributing a quote.
- Accuracy on real-world audio. Interviews happen in cafés and on doorsteps, not in studios. We only repeat what vendors themselves claim — and note where they hedge.
- Interview workflow. Some tools record the interview live; others only accept uploaded files after the fact. Those are different jobs.
- Price for a freelancer. A newsroom seat license and a per-story budget are very different constraints.
The 7 tools at a glance
| Tool | Best for | Free tier | Speaker labels | Where audio goes | Paid from |
|---|---|---|---|---|---|
| Otter.ai | Live transcription of calls | 300 min/mo (3 file imports ever) | Yes | US (AWS US West) | $8.49/mo annual |
| Good Tape | Source protection, EU privacy | See note — site shows conflicting free plans | Paid tiers | EU only | €20/mo (≈€16/mo annual) |
| Trint | Newsroom teams | None (7-day trial) | Yes | Your choice: EU or US | ~$52/seat/mo annual* |
| Rev | Accuracy on rough audio | 45 AI min/mo | Yes | US | $25.49/mo annual; humans $1.99/min |
| Descript | Audio/video journalists who publish | 60 min/mo | Yes | US cloud (region unstated) | $16/mo annual |
| Google Pinpoint | Document-heavy investigations | Free (full access on application) | Not documented | Google's cloud | Free |
| meaty (ours) | Recording the interview itself | 120 min/mo | Yes | EU | €4/mo |
All pricing and limits as of July 2026, from each vendor's site except where flagged (Trint's price is third-party-reported — see its section). Plans change constantly; check the vendor's pricing page before deciding.
The 7 tools, one by one
1. Otter.ai — best for live transcription of scheduled calls
Otter is the household name, and its strength is live transcription: it joins Zoom, Teams, and Meet calls and transcribes them in real time, and its iOS/Android apps record and transcribe in the field, with speaker identification on every tier (pricing, as of July 2026). The free plan includes 300 transcription minutes a month, capped at 30 minutes per conversation — but only 3 audio file imports, ever, which makes it nearly useless for uploading interview recordings. Pro is $8.49/user/month billed annually ($16.99 monthly).
On source protection: Otter stores data on AWS in the US (US West region) and holds SOC 2 Type 2 (security policies). It also states it trains its own models on customer conversations after de-identifying them (privacy & security page) — a posture some journalists will be fine with and others won't. Its own help docs name background noise as the biggest accuracy factor and note strong accents reduce quality (accuracy FAQ).
The honest caveat: as of July 2026, Otter faces a consolidated US class action alleging its meeting bot recorded participants without consent and that conversations were used for training (National Law Review). Those are allegations, not findings — but for source-sensitive work, they're worth knowing about. We compare Otter in more depth in our Otter.ai alternative guide.
2. Good Tape — best for source protection and EU privacy
Good Tape is the tool journalists recommend to each other for a reason: it was built inside Zetland, a Danish investigative newsroom, and it treats source protection as the product. All transcription is processed on servers inside the EU, it's ISO 27001 certified, it states plainly that it will never use your recordings to train models, and audio files are deleted automatically after transcription unless you choose to keep them (security page, as of July 2026). It transcribes 100+ languages (FAQ).
Pricing: Pro is €20/month billed monthly, or €192/year (≈€16/month) (pricing). The free tier is currently ambiguous on Good Tape's own site — its FAQ describes three 30-minute transcripts per month, while its free page describes a one-hour free trial — so check what you're offered at signup.
The honest caveat: Good Tape is upload-based. There's no live transcription and no meeting capture; even its mobile recorder app syncs audio for transcription afterward and can't record phone calls. Speaker labels are a paid-tier feature. It's the tool for the tape you already have, not for capturing the call.
3. Trint — best for newsroom teams
Trint was founded by Jeff Kofman, an Emmy-winning war correspondent, and it shows in the feature set: Trint Live streams a shared real-time transcript from a reporter's phone to the whole desk, every line is click-to-play verifiable against the audio before you publish, and the mobile app keeps recording through connection drops (live capture & verification). It transcribes 50+ languages and translates into 70+ (product page), with automatic speaker labels. Security is genuinely strong: you choose whether your data is stored in the EU or the US, it's ISO 27001 certified, and Trint states it never listens to your recordings to train its algorithms (security page, as of July 2026).
The honest caveat: the price, if you're a freelancer. There's no free plan (only a 7-day trial), and Trint doesn't publish a simple public price list — third-party coverage puts the Starter plan around $52/seat/month billed annually (~$80 monthly) for 7 files a month (Sonix's Trint pricing roundup, June 2026; we couldn't verify against a vendor page). Also note data is retained for the life of your subscription by default — deletion is on request, so source-protection housekeeping is on you.
4. Rev — best raw accuracy on rough audio
Rev's differentiator is that you can hand your hardest tape to a human. Its human transcription service starts at $1.99/minute with a 99%+ accuracy claim and ~12-hour turnaround, and Rev itself recommends humans over AI for accents and background noise (human transcription page, as of July 2026) — which is a usefully honest admission about AI's limits on messy field audio. The AI side offers a free 45 minutes/month (English only) and an Essentials plan at $25.49/seat/month billed annually (pricing), with speaker diarization on by default (Rev docs).
On data handling: Rev is a US company and says personal data may be stored in or accessed from the US (privacy policy). It holds SOC 2 Type II, and its transcriptionists work under NDAs (security page) — but with human transcription, people do hear your tape; that's inherent to the service. Rev states it doesn't sell customer data or allow it to train external LLMs (its wording).
The honest caveat: cost at scale. An hour-long interview transcribed by a human runs about $119 — fine for the make-or-break investigation quote, unsustainable for weekly reporting. And for EU source-protection standards, US jurisdiction is a real consideration.
5. Descript — best for audio and video journalists who publish
Descript is an editor first: you cut audio and video by editing the transcript text, which makes it the natural pick for podcast and video journalists who go from interview to published piece in one tool (descript.com). Speaker detection is automatic (help docs). The free plan includes 60 minutes of transcription a month; the Hobbyist plan is $16/user/month billed annually ($24 monthly) for 10 hours a month (pricing, as of July 2026).
Security posture is solid: SOC 2 Type II, AES-256 at rest, and sharing data to improve the service is opt-in and disabled by default (security page). Data lives on Amazon S3 or Google Cloud; Descript doesn't state a region, so assume US.
The honest caveat: transcription is the on-ramp, not the product. The hour caps are low if transcription is all you need, and language support is roughly two dozen Latin-alphabet languages — no Chinese, Japanese, or Russian, and one language per file (supported languages).
6. Google Pinpoint — best for document-heavy investigations
Pinpoint isn't a recorder — it's Google's free research tool for journalists, built to search enormous collections: PDFs, emails, images with OCR (including handwriting), and audio it transcribes in over 100 languages (about page; transcription languages). Upload an archive of interview files (up to 2 hours / 8 GB each) and every word becomes searchable (file types). It's free; full upload access requires applying as a journalist or academic researcher (access FAQ, as of July 2026).
Data handling is a mixed bag to read carefully: collections are private by default and private uploads aren't used to train LLMs, but Google's human reviewers may read a sample of your Pinpoint data (disconnected from your account by default), and usage data feeds Google product improvement (data usage page).
The honest caveat: it's a research tool, not a transcription workflow. There's no recording, no documented speaker labels for uploaded audio, approval is required for full access, and your investigation lives behind a Google account.
7. meaty — best for recording the interview itself (yes, it's ours)
Full disclosure: meaty is our product, so weigh this entry accordingly. We built it for the part of the job the six tools above mostly skip: capturing the interview as it happens. meaty records live from any phone browser — put the phone on the table for an in-person interview, no app-store install — or from the Chrome side-panel extension (published July 2026) for calls happening in your browser. No bot joins anything. When you stop, you get a speaker-labeled transcript plus a summary.
On the axes above: audio is processed and stored EU-hosted, the free tier is 120 minutes per month, and paid plans start at €4/month — deliberately priced for freelancers rather than newsroom seats. The workflow is capture-first: record on your phone, get the transcript, pull the quote. For the full method, see our guide to free interview transcription.
The honest caveat: meaty is built to record and transcribe interviews it captures — it's not a bulk processor for an archive of old tape (Pinpoint or Good Tape fit that better), and there's no human-verbatim option for legally sensitive transcripts (that's Rev). If those are your jobs, pick those tools.
Which tool should you pick?
Match the tool to the job, not the brand to the hype:
- You work with vulnerable sources in Europe: Good Tape for uploaded tape, meaty for recording live — both EU-hosted with no-training policies. Trint if the newsroom pays and you select EU storage.
- You're a freelancer watching every euro: meaty (free 120 min/month, from €4/month) or Rev's free AI tier for short English clips. Trint's seat pricing isn't built for you.
- Your tape is rough — accents, wind, a crowded café: budget for Rev's human transcription on the stories that justify $1.99/minute; every AI tool here degrades on noisy audio, and Otter's own docs say as much.
- You live in scheduled video calls: Otter, with eyes open about US hosting, model training, and the pending litigation.
- You publish audio or video: Descript, since editing by transcript collapses two tools into one.
- You inherited 40 hours of tape and a document dump: Google Pinpoint, if you qualify for full access.
Frequently asked questions
What is the best free transcription tool for journalists?
For recording interviews, meaty gives you 120 free minutes a month with speaker labels and EU hosting. For transcribing an existing archive, Google Pinpoint is free for approved journalists. Otter's free tier offers 300 live minutes a month but only 3 file uploads ever, which rules it out for transcribing recordings you already have. All limits as of July 2026.
Which transcription tools keep interview audio in the EU?
Good Tape processes everything on EU servers by design, and meaty is EU-hosted. Trint lets you choose EU or US storage on its AWS infrastructure. Otter, Rev, and Descript store data in the US (or don't publish a region), which matters if your source-protection standard — or your editor's lawyer — requires keeping audio under EU jurisdiction.
How do I protect a source when using AI transcription?
Check three things on the vendor's security page before uploading: where the audio is physically processed and stored, whether the company trains AI models on your recordings, and how deletion works. Prefer tools that auto-delete audio or let you delete it immediately, and remember that human transcription services mean real people hear your tape, NDA or not.
What is the most accurate way to transcribe a noisy interview?
Human transcription — Rev claims 99%+ accuracy and itself recommends humans over AI for accents and background noise, at $1.99/minute as of July 2026. If that's out of budget, improve the capture instead: put the recorder between you and your subject, pick the quiet corner, and state names at the start so speaker labels anchor cleanly.
The bottom line
The winners by job: Good Tape for source-sensitive uploaded tape, Rev when accuracy is worth paying humans for, Trint for funded newsrooms, Descript for publishing audio and video, Pinpoint for investigations, Otter for a life of scheduled calls — and meaty for the moment none of them cover: actually recording the interview, from the phone already in your pocket, EU-hosted, free to start.
Try meaty free — record your next interview from any phone browser and get a speaker-labeled transcript and summary the moment you stop.