5 Best Cantonese Speech-to-Text Tools in 2026: Free Tiers and Mixed Chinese-English Compared

Why do you need Cantonese audio transcription?

Ask yourself a practical question: do you want to "turn audio into text" or "turn meetings into workable data"? The former depends on transcription accuracy and free quota, while the latter depends on summaries, action items, exports, and team collaboration.

For Cantonese meetings, interviews, classes, and phone recordings, the most troublesome part is often not recording but re-listening afterward. The original article mentions that manually transcribing a 60-minute recording takes an average of 3 to 4 hours. That's why more and more people are looking for online platforms to convert audio files directly into transcripts.

This article covers 5 common options: Tinrec, Subanana, cSubtitle, Google AI Studio (Gemini), and SoundType AI. I'll first explain who each is for, then show you how to choose using three criteria: free quota, language handling, and output format.

Option 1 | Tinrec — An AI meeting notes and team collaboration assistant with a relatively complete post-meeting workflow

Tinrec is an AI meeting notes and collaboration tool for individuals and teams. It doesn't just turn audio into text; it organizes meetings into searchable, summarizable, queryable, and exportable data. If what you care about is "how much time you'll spend organizing after the meeting," it focuses on this the most in this list.

The first practical point is recording online meetings without a bot. The desktop version directly captures computer system audio and can handle Zoom, Google Meet, Microsoft Teams, and Webex without needing an extra meeting bot to join. For teams that don't want another participant, this is very convenient.

The second is post-transcription organization. It generates summaries, chapters, key points, and action items, and can also do AI Q&A around meeting content. When you need to deliver, it can generate reports, tables, and documents, and export to Notion, Google Docs, OneNote, Dropbox, and other tools.

The third is Chinese and mixed Chinese-English. Tinrec covers Chinese, Traditional Chinese, and multilingual meeting scenarios, supports real-time translation and bilingual viewing, and is suitable for meetings that are mainly in Chinese with English terms mixed in. Actual performance will still be affected by audio quality, accent, noise, and technical terminology.

The team version has independent team spaces, member roles, seats, shared import quota, usage analytics, audit logs, and an audio recycle bin; when members leave, team resources can be handed over. Note that it is currently team-level sharing, not file-level ACL.

For pricing, the personal version has a free version, weekly card, Pro monthly card, and Pro annual card; actual prices are subject to the official website. The team version offers a 7-day trial with 1 free seat and 300 minutes of team shared import quota for first-time eligible teams; monthly payment is USD 29.80 per paid seat per month, annual payment is USD 199 per paid seat per year, about USD 16.58 per month; each paid seat gets 2,000 minutes of team shared import quota per month. Real-time recording is currently not deducted by minutes; file and web imports use team shared minutes.

Who it's for: office workers, project teams, and managers who mainly work in Chinese and often need post-meeting organization and team sharing. If you only need to convert pure Cantonese colloquial speech into written Chinese, Subanana's positioning is more direct; Tinrec's strength is the post-meeting workflow.

Option 2 | Subanana — A specialized choice for Cantonese transcription and colloquial/written output

Subanana is a speech-to-text platform for Cantonese. Upload audio files, videos, or paste YouTube links, and the AI will convert Cantonese conversations into transcripts with speaker separation and punctuation. It explicitly supports mixed Chinese-English, and output can be selected as colloquial or written, Traditional or Simplified, which is very suitable for Hong Kong meetings and interviews.

Free accounts can upload and preview the first 15 minutes of transcription per file without a credit card; exporting files is a paid feature. Suitable for those who need Cantonese transcripts, subtitles, or colloquial-to-written conversion.

Option 3 | cSubtitle — An online video and audio transcription platform without account registration

cSubtitle is an online video and audio transcription platform. Its biggest feature is that you can upload without registering an account. It offers Cantonese (HK) and English options, can handle up to 4GB and 5 hours per file, and supports MP4, MOV, MP3, WAV, M4A, OGG, and WMA.

Output includes text files with punctuation and paragraphs, Word files with time codes, and also generates .sub and .vtt subtitle files. The free version has a quick 3-minute limit, suitable for trying short videos first. Suitable for those who need subtitle files temporarily and don't want to create an account.

Option 4 | Google AI Studio (Gemini) — The free Cantonese transcription workflow mentioned by the original author

Gemini on Google AI Studio is not a dedicated transcription app but a general AI tool. The original author's experience is: free, friendly to Cantonese, can handle long content, and can directly ask Gemini to further analyze, organize, and summarize the transcript; usable on desktop and mobile.

Note that this is a personal workflow; the operational threshold and free quota are still subject to Google's official terms, and it does not have team seats or auditing. Suitable for those with zero budget who are willing to organize things themselves.

Option 5 | SoundType AI — A mobile transcript and AI summary tool

SoundType AI is a mobile transcript and speech-to-text tool. Its store page advertises transcripts, audio file to text, AI summaries, and speaker identification, supporting over 90 languages.

Some user reviews on the store page say Cantonese conversion is about 85–90% usable but not good enough; this is user opinion, not official data. Some also mention that the 8-minute trial is rather short. Prices and plans are subject to the app page. Suitable for those who record casually on mobile and need summaries and speaker separation.

With so many options, how should you choose?

• Mainly Chinese with frequent English terms, need post-meeting summaries, action items, Q&A, and export → Tinrec. Its strength is completing the post-transcription work together. • Need team sharing of meeting data, managing member seats, viewing usage and auditing → Tinrec team version. This is one of the few options in the list that clearly explains governance features. • Want bot-free recording of Zoom, Meet, Teams, Webex → Tinrec desktop version. The premise is that you actively start recording, not automatically join all calendar meetings. • Need Cantonese colloquial to written conversion, Traditional/Simplified output, and subtitles → Subanana. • Temporary short videos, no account, need .sub/.vtt subtitles → cSubtitle. • Zero budget, willing to operate AI yourself → Google AI Studio; for casual mobile recording, consider SoundType AI.

5 things to know before using Cantonese audio transcription

1. Check the free quota first: cSubtitle free 3 minutes, Subanana free only previews the first 15 minutes per file and export is paid, SoundType trial is rather short; Tinrec personal version has a free version, team version trial is 7 days with 300 minutes. You can try short meetings first; for long meetings, calculate the cost.
2. Mixed Chinese-English and output style: Subanana can choose colloquial or written; Tinrec can organize Chinese and mixed Chinese-English. If you want the original transcript, first confirm whether the output is colloquial or written.
3. Output format determines subsequent workload: if you need Word with time codes and subtitle files, look at cSubtitle; if you need summaries, action items, Q&A, and multi-format export, look at Tinrec.
4. Team permissions are not file-level ACL: Tinrec team version mainly uses team-level sharing; roles control governance, seats control usage; members without seats can view, play, and read but cannot upload, edit, or export. Don't write "can view" as "can edit."
5. Get consent before recording: when recording and transcribing meetings, comply with local regulations and obtain participant consent when necessary.

FAQ

Is there a completely free Cantonese audio transcription platform?

You can start with cSubtitle's free 3 minutes, Subanana's first 15 minutes preview per file, or the free Google AI Studio workflow mentioned in the original article; but for long-term and long meetings, you'll usually hit quota or export limits. If you're just starting out, Tinrec's free version lets you experience the basic quota.

Which one is best for mixed Chinese-English?

Subanana explicitly supports mixed Chinese-English and colloquial/written output; Tinrec focuses on Chinese meetings and mixed Chinese-English content organization. Actual performance will be affected by accent, noise, and terminology; it's recommended to try with your own recordings first.

Can Tinrec directly transcribe Cantonese?

Tinrec's product description emphasizes Chinese, Traditional Chinese, and multilingual meeting scenarios, and supports mixed Chinese-English organization, but it does not publicly promise a fixed Cantonese accuracy rate. If your content is a lot of colloquial Cantonese and you need written conversion, Subanana's positioning is more direct; if it's general Chinese meetings and post-meeting organization, Tinrec is more suitable.

Is the free version enough?

For short meetings and occasional recording organization, yes. If you have meetings every week and need export, summaries, action items, or team sharing, the quota and features of paid plans will be more practical. Prices and benefits are subject to each platform's purchase page.

How is Tinrec team version priced?

The first eligible team trial is 7 days, 1 free seat, and 300 minutes of team shared import quota. Team monthly payment is USD 29.80 per paid seat per month; annual payment is USD 199 per paid seat per year, about USD 16.58 per month. Each paid seat gets 2,000 minutes of team shared import quota per month; real-time recording is currently not deducted by minutes. Actual prices are subject to the official purchase page.

Conclusion

There is no single answer for Cantonese audio transcription; the key is how much time you still need to spend after finishing the transcript. If you need Chinese meetings, mixed Chinese-English, post-meeting summaries and action items, and team knowledge retention, Tinrec is the one in this list with relatively complete feature coverage and is most suitable for continuing work.

If you focus on Cantonese colloquial to written conversion, Subanana is more targeted; for temporary short videos and subtitles, you can try cSubtitle first. With a limited budget, run your own recording through the free quota first, then decide whether to pay.

References

• Online AI speech, video, audio file auto transcription - free trial - cSubtitle
• Free Cantonese speech-to-text app 2026 | Subanana
• "Free + Cantonese OK: How to use Google AI Studio to turn your recording into..."
• 2026 Latest 4 Cantonese audio transcription tools recommendation: solving mixed Chinese-English and meeting notes pain points - Tinrec Blog
• SoundType AI ultra-accurate transcript, speech-to-text, AI dictation - Google Play App