Accurate in 90+ languages
It detects the language for you — English, Afrikaans, Spanish, Zulu, Japanese and many more. Even mixed in one recording.
90+ languages · Afrikaans included
Upload a file or paste a YouTube link. Get an accurate, timestamped transcript in minutes — with speakers, subtitles and an AI summary when you want one.
90+ languages, including Afrikaans — with the same transcription quality on every plan, including Free.
01Try it
Search it, switch to the AI summary, click a timestamp. This is exactly what you get.
02How it works
Any audio or video file — MP3, M4A, WAV, MP4, MOV and more — or a YouTube link.
Parlova detects the language and transcribes every word, with times and speakers.
Copy the text, find any word, or download it as subtitles, Word or PDF.
03Features
Every plan gets the same transcription quality. Plans differ in how much you can do.
It detects the language for you — English, Afrikaans, Spanish, Zulu, Japanese and many more. Even mixed in one recording.
Speakers are separated, and you can name them once and reuse the names.
A summary, key points, decisions and action items for every recording.
Download as text, subtitles for video, a spreadsheet, Word or PDF.
Teach it names, places and product words, up to 50 of them.
Uploaded audio is deleted once it's transcribed. Our speech providers don't store it or train on it.
A REST API, a command line tool and an MCP server, so your scripts and AI apps can transcribe. Plus and Pro.
04Pricing
Start free and upgrade any time. Already have an account?
$0one-time allowance
For trying it out
$5/ month
Or $50 a year — 2 months free
$10/ month
Or $100 a year — 2 months free
Prices in US dollars. Paid plans renew automatically until cancelled. Payments are non-refundable except as stated in our refund policy or required by law. Review our Terms and Privacy Policy before buying.
05FAQ
Parlova is an online audio and video transcription service. Upload a recording, record in the browser or paste an accessible YouTube link to get text you can search, copy and download. It supports 90+ languages, including Afrikaans. Speaker labels and subtitles need Plus or Pro; AI summaries need Pro.
Free includes 30 minutes total per account as a one-time lifetime allowance, with up to 5 minutes per day and 5 minutes per recording. Files can be up to 15 MB. No card is needed. The allowance does not reset monthly or after deleting recordings. Compare the plans.
More than 90, including English, Afrikaans, Dutch, German, French, Spanish, Portuguese, isiZulu, Japanese, Chinese, Hindi and Arabic. Parlova detects the language by itself, or you can choose it.
Almost any audio or video file: MP3, M4A, WAV, OGG, FLAC, AAC, MP4, MOV, MKV, WEBM and more. You can also paste a YouTube link instead of uploading.
Up to 5 minutes on Free, 1 hour on Plus and 3 hours on Pro. Free includes 30 minutes in total as a one-time allowance. Plus includes 10 hours each month and Pro includes 30 hours each month. Free is limited to 5 minutes per day and shared daily capacity; the daily allowance resets at 00:00 UTC.
You keep your plan until the end of the period you've paid for, and you won't be charged again. After that you move to Free. You can cancel in two clicks, any time.
Payments aren't refunded for partly used periods, because you keep access until the period ends. Mistakes like double charges are always refunded, and your legal rights always apply. Read the refund policy.
Yes. Your audio is only used to make your transcript, by providers that don't store it or train on it. Uploaded audio is deleted once it's transcribed, and you can delete any transcript at any time. Read the privacy policy.
It joins two old roots. Parl- is the word for speaking in Italian (parlare) and French (parler) — the same root as “parley” and “parliament”. Nova is Latin for “new”. Together: speech made new — your spoken words turned into something you can read, search and use.
Yes — Plus and Pro include a REST API with access tokens, a command line tool, and an MCP server that lets apps like Claude and Cursor transcribe files for you.
Parlova · Speech made new
30 minutes total on Free, up to 5 minutes/day. Your free allowance does not reset. No card needed.
Welcome back
New transcripts and summaries are unavailable right now. Please check back shortly.
You can still open and download your existing transcripts.
Choose light or dark, or follow your device.
—
Minutes of transcription used each month.
| Month | Recordings | Minutes used |
|---|
Every finished transcription, newest first — including ones you've since deleted.
Your terminal is showing a sign-in code. Make sure it matches this one:
····-····
Only approve if you just ran parlova login yourself and the code matches. Approving lets that computer — and the AI apps you connect to it — transcribe with your account.
The CLI and MCP are included with Plus and Pro.
Transcribe from your own scripts, apps, terminal and AI assistants. Upgrade to Plus or Pro to create an access token — this page shows what you can do.
Use a token to transcribe from your own code. Send it in the Authorization: Bearer … header. Tokens can upload, check and download transcriptions, but can't change your account, plan or password. Keep them secret — anyone with a token can use your minutes.
Copy your new token now. For your security it won't be shown again.
Transcribe from your own scripts, apps, terminal and AI assistants. Upgrade to Plus or Pro to create an access token — this page shows what you can do.
You'll need an access token — create one under Access tokens.
status is completed.| Request | What it does |
|---|---|
POST /api/jobs/upload?filename=… | Upload one file as the raw body. Optional: language (e.g. en, default auto), diarize=true (who said what), denoise. Returns a list with the new job. |
POST /api/jobs | Transcribe YouTube links. JSON body: {"urls": ["…"], "language": "auto", "diarize": false}. |
GET /api/jobs | Your recent jobs. |
GET /api/jobs/{id} | One job: status (queued, running, completed, failed, cancelled), progress, text, paragraphs with timestamps and speakers, and error if it failed. |
GET /api/jobs/{id}/file/{format} | Download as txt, srt, vtt, csv, md, json, docx or pdf. |
POST /api/jobs/{id}/cancel | Stop a job that's waiting or running. |
DELETE /api/jobs/{id} | Delete a job and its transcript. |
GET /api/usage | Minutes used this month and your usage history. |
Errors come back as JSON: {"error": "…"} with a 4xx status — for example 402 when your monthly minutes run out, 413 for a file that's too big, and 429 when you have too many jobs waiting. API use counts towards your plan's minutes.
Transcribe from your own scripts, apps, terminal and AI assistants. Upgrade to Plus or Pro to create an access token — this page shows what you can do.
Transcribe from your terminal with the parlova command. It's one small file and needs Python 3.8 or newer.
1. Install
2. Sign in
3. Transcribe
Run parlova --help for every option. Prefer not to run an installer? Download the file and run it with python3 parlova.py.
On a server without a browser, run parlova login --no-browser and open the link on any device — or use parlova login --token with a token from Access tokens. Your sign-in is saved in one file only you can read; parlova logout removes it.
Transcribe from your own scripts, apps, terminal and AI assistants. Upgrade to Plus or Pro to create an access token — this page shows what you can do.
Let Claude, Cursor and other AI apps transcribe for you. The Parlova CLI includes an MCP server (parlova mcp) that runs on your computer, so the AI can transcribe files on your computer as well as YouTube links. Install the CLI and run parlova login first — that signs it in through your browser. Then add it to your app. No token goes in your app's settings.
It prints something like this — paste it in:
Using several Parlova accounts, or running on a server? Set PARLOVA_TOKEN in the MCP server's environment instead of signing in.
Tools your AI app gets
transcribe_file — transcribe an audio or video file on your computertranscribe_youtube — transcribe a YouTube linkget_transcript — fetch a transcript as text, or as srt, vtt, csv, md or jsonlist_transcripts — your recent transcriptionsget_usage — minutes used this monthTry asking: “Transcribe ~/Downloads/team-meeting.m4a with speakers and list the action items.”