5 Best Cantonese Speech-to-Text Tools in 2026: Which Free Tier Actually Works?

Last semester, I was in a group with two exchange students from Hong Kong for our graduation project. Midway through discussions, they would naturally switch back to Cantonese. I recorded the audio to organize later—but after putting on headphones, I gave up within three minutes; I couldn't keep up when they spoke fast. During that time, I tried almost every Cantonese speech-to-text tool I could find: some claimed to be free, but after transcription, you'd find they only let you hear a short segment; some only recognized Mandarin, and when you fed them Cantonese, the output was gibberish; others required applying for an API key and setting up for ages before you could start. This article shares my hands-on experience this semester. The tool I ended up keeping is Tinrec, and I'm still using it even after finishing the project.

What Can Cantonese Speech-to-Text Do for You?

Scenario 1: As soon as a group discussion ends, the transcript and conclusions come out together Previously, our group would discuss for two hours, and after the meeting, someone would spend another hour organizing who said what and what the conclusions were. Now I put my phone on the table and hit record. With Tinrec, by the time the discussion ends, the transcript is already formed, and the AI summary has listed conclusions and to-dos—who needs to submit what and by when, all clear at a glance.

Scenario 2: Cantonese lectures or talks—no fear of not finishing your notes Some general education courses invite Hong Kong speakers who write little on the board and speak quickly. Using Tinrec to record and transcribe simultaneously, I can listen and read the text at the same time, confirming on the spot whether I missed any key points, without regretting it after class.

Scenario 3: Interview recordings for graduation projects After an hour-long interview, the most painful part is transcribing word by word. Upload the audio or video file to Tinrec, and after transcription, you can even ask it questions directly—much faster than listening to it over and over myself.

Scenario 4: Organizing Cantonese video materials When citing Cantonese video content in reports, save the audio and upload it to turn it into searchable text, without constantly pausing to copy subtitles.

What Are the Core Capabilities of Cantonese Speech-to-Text?

Real-time recording and transcription: text appears as you speak It's not about waiting until after recording. Tinrec supports real-time transcription, so you can watch the text on the spot during discussions. If you notice a sentence wasn't transcribed, you can immediately ask your classmate to repeat it—super useful for group discussions.

Mixed Chinese-English and bilingual comparison Hong Kong classmates often mix Chinese and English in one sentence, which was where I initially stumbled—models that only handle Mandarin would force English words into phonetically similar Chinese. Tinrec covers Chinese, Traditional Chinese, and multilingual meeting scenarios, and supports real-time translation, allowing you to compare original text, translation, or bilingual views. For someone with average Cantonese listening skills like me, this is very helpful.

AI summary, chapters, and key points The most troublesome part of long meetings is reviewing. Tinrec automatically generates summaries and splits long recordings into chapters; when I need to find the "budget discussion" section, I can jump directly to the chapter without listening from the beginning.

To-do extraction and AI Q&A After transcription, I don't have to dig through it myself. I can ask Tinrec directly, for example, "What was the group's final conclusion on the topic scope?" It answers based on semantics; to-dos are also organized separately, which saved me a lot of time during the project crunch.

Multi-format export and team hotwords Transcripts can be exported in common file formats and taken to Notion, Google Docs, OneNote, or Dropbox for further editing. The team version also allows maintaining team hotwords, adding proper nouns and common terms, which is quite helpful for projects with lots of technical jargon like ours.

Real-World Application Scenarios

Scenario 1: Project discussions with Hong Kong group members Our group discussed twice a week, two hours each time. Previously, we had to schedule an extra hour after meetings to organize notes. Now I record with Tinrec at the start of discussions, check the AI summary to confirm conclusions, then go back to the transcript to fill in details. Over the semester, the time saved on organizing alone was enough for me to prepare for two more final exams.

Scenario 2: Cantonese lectures and online talks Remote lectures often lack subtitles, and speakers talk fast. I use Tinrec to listen and read real-time text, mark key points on the spot, and review with AI summaries after class—much more reliable than borrowing a classmate's notes.

Scenario 3: Graduation projects and research interviews Interview recordings often last an hour or more, and interviewees frequently mix Chinese and English. After uploading audio to Tinrec for transcription, I directly ask, "How many times did the interviewee mention price-related concerns?" The answer comes much faster than searching myself, and I can cross-reference the original text when citing.

Scenario 4: Internship interviews and cross-regional meetings During my internship, when having online meetings with Hong Kong contacts, I dreaded an unknown account appearing in the meeting room. Tinrec's desktop version directly captures the computer's system audio without inviting a meeting bot, so the other party won't see an extra participant—a big plus for client meetings.

5 Things to Consider When Choosing a Cantonese Speech-to-Text Tool

1. Is free really free, or does it only let you hear the first few minutes? Many tools attract you to upload with "free," but after transcription, they only open a short transcript; to copy the full text, you have to pay. I test with a ten-minute Cantonese recording first to see how far the free quota goes. Tinrec's free version offers a basic quota to experience recording, importing, summaries, and transcription. When the quota runs out, you can choose a weekly, monthly, or yearly plan as needed—no "transcribe then find out you need to pay to unlock" situations.

2. Does the model truly handle Cantonese and mixed Chinese-English? Some tools are primarily trained on Mandarin data and start guessing when encountering Cantonese or mixed Chinese-English. Rather than looking at promotional numbers, test with your own recordings. Tinrec covers Chinese and multilingual meeting scenarios, but actual results are still affected by recording quality, accent, environmental noise, and overlapping speakers. I still cross-check important content against the original.

3. Can it help you organize after transcription? Transcripts are just the first step. Whether it can automatically generate summaries, chapters, to-dos, and whether you can continue asking questions about the content—that's the key to saving time. Tinrec treats this as a main selling point; after transcription, I hardly need to organize manually.

4. Do the output formats fit your workflow? I use Google Docs for reports and Notion for notes. If transcripts can only be downloaded in one format, I'd still have to copy and paste. Tinrec supports multiple export formats and can directly integrate with Notion, Google Docs, OneNote, and Dropbox—very practical for student reports.

5. Data ownership and group sharing Personal version data belongs to you, but group projects often require sharing. Tinrec's team version is a separate team space where meeting data belongs to the team. You can manage members, roles, and seats, plus usage analytics and audit logs. Members without a seat can still view, play, and read data, but need a seat to record, upload, and edit. Pricing: first eligible team trial is 7 days, 1 free seat, 300 minutes of team shared import quota; paid seats are USD 29.80 per month, or USD 199 annually (about USD 16.58/month), with each paid seat providing 2,000 minutes of team shared import quota per month. Real-time recording under active team plans and seats is currently not deducted by minutes. For student groups, unless it's a long-term shared project, using the free version individually is sufficient.

Frequently Asked Questions

Q1: Is there truly completely free Cantonese speech-to-text? Completely free usually comes with conditions, such as limiting file length, only previewing part of the content, or occasional queuing. A more practical approach is to find a tool with a free quota and try it. Tinrec's free version offers a basic quota to experience recording, importing, summaries, and transcription. If you only organize one or two recordings a week, it's usually enough.

Q2: Can mixed Chinese-English Cantonese speech be transcribed? Yes, but it depends on whether the tool truly supports mixed languages. From my experience with Tinrec, English words are kept as original and not forced into phonetically similar Chinese. However, if there are many proper nouns, it's advisable to build a terminology list using team hotwords first.

Q3: What's the approximate accuracy rate? I don't recommend looking at a single number. The same recording can yield very different results in a quiet classroom versus a noisy café; accent, speed, and overlapping speakers also affect it. Test with the type of recording you encounter most. Tinrec provides high-quality transcription for clear audio and common scenarios, but important information should still be cross-checked against the original.

Q4: Can I upload files recorded on my phone? Yes. Tinrec supports mobile recording and uploading existing audio or video files, whether class recordings or interview files.

Q5: How can a group share the same transcript? The personal version is for individual use; for sharing, the team version is recommended. Data in the team space belongs to the team, and members can view, play, and read. If a member leaves or loses their seat, team resources are transferred to other active members and won't disappear with the individual account.

Q6: What should I note before recording? First, check local regulations and obtain consent from those present if necessary. For campus interviews, I also explain the purpose before recording—that's basic courtesy.

Conclusion

Cantonese speech-to-text in 2026 is no longer a question of "is there a tool" but "is the free quota enough, can it transcribe mixed Chinese-English, and can it help you organize after transcription." My own order is: first test a segment of your own Cantonese recording with the free quota, then confirm whether it has summaries, to-dos, and Q&A, and finally look at the price. This semester, from group discussions to Cantonese lectures to graduation project interviews, I've relied on Tinrec. If you also often get stuck at "recorded but no energy to organize," start with Tinrec's free version, throw in the most troublesome Cantonese recording you have, and see if the output is something you can use directly.

References

• Free Cantonese Speech-to-Text App 2026 | Subanana
• Online AI Speech, Video, Audio File Auto Transcription - Free Trial - cSubtitle
• "Free + Cantonese OK: How to Use Google AI Studio to Transcribe Your Recordings" Finding a good transcript app is painful; I don't know if you've tried different recording-to-text apps—either fake free, charging a lot after the first few minutes, or only supporting English, not Cantonese, or requiring a lengthy API setup. Recently I found Google Gemini is world-class in this area: - [Truly Free] No subscription fee after a few uses - [Cantonese-friendly] Gemini accurate
• Tested 4 Best "Speech-to-Text" Online Tools, Free Instant Audio to Transcript
• AI Speech-to-Text Fast Audio File Online Transcription, Transcripts and Video Subtitles - Free Trial - cSubtitle