Cockatoo transcribes audio and video in 90+ languages with speaker labels and exports. Free plan; Pro from $11.99/month billed annually. Full review.
Researched and fact-checked by Bartu Cavusoglu · Updated October 26, 2026
Cockatoo is an AI transcription service that converts audio and video recordings into text. It runs in the browser, supports more than 90 languages and dialects, and produces transcripts with speaker labels and timestamps. The company states accuracy of up to 99.8% on clear audio.
Besides transcription, Cockatoo translates documents such as PDF, Word, PowerPoint, Excel and text files, and includes an AI text editor called Canvas. An iOS app, Voice Notes, lets users record and transcribe on a phone.
The product targets people who work with recordings regularly. Customer reviews on the site mention interviews, podcasts, dissertation research, meeting minutes and multilingual conversations.
Users sign up for a free account, then upload a file by drag and drop or the file picker. In the uploader they choose a language and a transcription model, and can enable speaker detection. Progress appears in a section called My Drive, which also shows the status of each file when several are uploaded.
Two modes are available. V4-Fast is built for turnaround and runs at about 120x real time, so a one-hour recording takes about 30 seconds. V4-Thinking aims for higher accuracy and runs at about 20x real time, about 3 minutes per hour. V4-Thinking has a weekly usage limit.
Once a transcript is finished, it opens in an editor where the recording can be played alongside the text. Users correct wording, rename speakers and export the result.
The editor is the center of the workflow. Users can edit words or sentences, jump to any timestamp, reassign text to another speaker, and switch between speaker, paragraph and sentence views. Exports cover PDF, DOCX, TXT and SRT subtitles, with timestamps and speaker labels preserved.
Input support is broad. The help center lists dozens of audio extensions, including MP3, WAV, M4A, AAC, FLAC and OGG, and video extensions such as MP4, MOV, MKV, AVI and WEBM. Video files need an audio track, and raw PCM files must be exported as WAV first.
Paid plans add unlimited file translation, 2 TB of encrypted storage and unlimited Canvas chats and edits. Team plans add shared files and centralized billing.
Cockatoo is an AI transcription service that converts audio and video files into text in more than 90 languages. It adds speaker labels and timestamps, includes an editor, and exports to DOCX, PDF, TXT and SRT.
The Free plan costs $0. Pro is listed at $11.99/month billed annually ($143.88 per year), and Team at $20.97/month billed annually for 3 people, or $6.99 per person per month. The homepage shows Pro at $29.99/month, which appears to be the monthly rate.
Yes. The free account includes 3 transcripts per day with no credit card required. It covers 90+ languages, speaker labels and timestamps, with limited access to transcription and translation features.
Cockatoo states accuracy of up to 99.8% on clear audio. Background noise, heavy accents and overlapping speakers lower it, and the editor lets users correct the text. The V4-Thinking mode is intended for higher accuracy.
V4-Fast transcribes at roughly 120x real time, about 30 seconds for a one-hour recording. V4-Thinking runs at roughly 20x real time, about 3 minutes per hour. Upload time and demand can add to the wait.
It accepts many audio formats such as MP3, WAV, M4A, AAC, FLAC and OGG, and video formats such as MP4, MOV, MKV, AVI and WEBM. Video files must contain an audio track. Transcripts export as DOCX, PDF, TXT and SRT.
Yes. Transcripts include speaker labels when speaker detection is enabled. Speakers can be renamed or reassigned in the editor.
Yes. Cockatoo Voice Notes for iOS lets users record voice notes and transcribe them. The website can also be accessed from a browser.