TinrecAI Meeting Recorder

4 Free Cantonese Speech-to-Text Tools in 2026: Free Limits and Transcript Uses

A roundup of 4 free speech-to-text options that handle Cantonese, comparing free quotas, spoken-to-written conversion, speaker labeling, and subtitle export formats, plus common pitfalls and a method to verify with your own recordings. Information verified as of August 2026.

Productivity Tips
QING
November 17, 2026
72 min
1 views

Turn recordings into transcripts and summaries in minutes

Upload audio or video for multilingual transcription, AI notes, and action items

4 Free Cantonese Speech-to-Text Tools in 2026: Free Limits and Transcript Uses

Many people think that if they just drop a Cantonese recording into a tool advertised as "free" speech-to-text, a transcript will magically appear.

But after a few tries, you'll find that the real obstacle isn't price—it's these things: models that understand Mandarin can't understand "係咪" (is it?), "唔該" (please/thanks), or "搞掂" (done); the free quota only covers the first 15 minutes; and the output is all colloquial, so you still have to rewrite it into written Chinese afterward.

This article rounds up 4 free Cantonese transcription options that work in the US, clearly explaining the free boundaries, output formats, and suitable scenarios, and ends with a method to "verify with your own recording."

Bottom line first: whether a free option works depends on whether your task is to "quickly understand the content" or to "deliver a usable transcript or subtitles." These two tasks need different tools.

The four most common misconceptions about free Cantonese transcription

Misconception 1: Treating "supports Chinese" as "supports Cantonese." Many tools only have "Chinese" in the language menu, which is actually a Mandarin model forced to fit. Cantonese has different vocabulary, word order, and sentence-final particles. Forcing it will produce transcripts full of sentences where you recognize the characters but can't understand the meaning.

Misconception 2: Looking only at the officially claimed accuracy rate. The high accuracy rates advertised by vendors are mostly tested in quiet environments. Real meetings have background noise, people talking over each other, and mixed Chinese-English, so performance will be noticeably different. Always test with your own actual recordings.

Misconception 3: Thinking the transcript is the end goal. In spoken Cantonese, people say "係咪" (is it?), "唔該" (please/thanks), "搞掂" (done). Formal records should be written as "是否" (whether), "謝謝" (thank you), "完成" (completed). If the tool only "types what it hears," the time you save on typing will all be spent on editing.

Misconception 4: Treating the free quota as permanently available. Free plans are usually "first N minutes per file" or "N minutes per month." Running out at the end of the month is a hassle. Estimate how long you'll record in a month before deciding whether to upgrade.

Several factors that affect transcription results

From hitting upload to getting a usable transcript, there are actually several steps:

  • Upload method: Uploading an audio file, pasting a YouTube link, or recording directly in the browser.
  • Language model: Is there a dedicated Cantonese model, or does it share a Mandarin model?
  • Speaker separation and punctuation: Does it automatically distinguish who is speaking in multi-person conversations, and does it automatically add punctuation?
  • Spoken or written output: Can it directly output written Chinese, or only transcribe verbatim?
  • Export formats: Plain text, Word, or subtitle files like SRT and VTT.

If any one of these steps doesn't work well, the word "free" becomes meaningless. So below, I won't use "free" as the sole criterion, but will mark the boundaries item by item.

The 4 tools compared in this article

  • Subanana: Focuses on Cantonese speech-to-text. You can upload audio or video files, paste YouTube links, choose between spoken and written Chinese, and preview the first 15 minutes of each file for free.
  • cSubtitle: Primarily Chinese-focused, with "Cantonese (HK)" in the language menu. You can upload and process without an account, but the free version has limitations.
  • YouTuber AI: A free tool. Upload a file or a YouTuber link to transcribe. The official claim is up to 96% accuracy.
  • Tinrec: Positioned as an AI meeting notes and team collaboration tool. The free version offers a basic quota, with a focus on Chinese and mixed Chinese-English content; for pure Cantonese spoken content, it's recommended to verify with your own audio first.

One thing to clarify upfront: This article is based on public information verification and feature comparison, not a controlled timed test. Therefore, there won't be numbers like "transcribed in seconds." Prices and free quotas are based on official pages as of August 2026.

Comparison method and conditions

Since I don't have traceable, uniform timing records, it wouldn't be honest to present others' speed as my own test results. So here I compare using "verifiable conditions": free scope, language processing, speaker and punctuation, export formats, and suitable tasks.

If you want to run your own test, here's a suggested approach:

  1. Prepare a Cantonese recording of about 10 minutes, with 2 speakers and background noise, preferably including a few mixed Chinese-English terms.
  2. Use the same file, same computer, and same network environment for all four tools.
  3. Start timing when you hit start transcription, and stop when the transcript appears on screen.
  4. After transcription, check three things: whether mixed Chinese-English terms are correct, whether names and numbers are correct, and whether speakers are separated.
  5. Test at least twice to avoid occasional queuing or failures.

Comparison summary table of the four tools

Tool Free conditions (per official announcements) Cantonese and spoken language handling Speaker and punctuation Export formats Who it's for
Subanana Free preview of first 15 minutes per file Dedicated Cantonese model, spoken or written Chinese selectable, traditional or simplified output Automatic speaker separation, automatic punctuation SRT, VTT, TXT, DOCX, XLSX, Markdown People who need to deliver transcripts or subtitles
cSubtitle Free version available, actual quota per official site; no account required Has "Cantonese (HK)" option, Chinese-oriented Output includes punctuation and paragraphing Word file (with time codes), .sub, .vtt subtitle files People who only need subtitle files and don't want to register
YouTuber AI Free tool, quota details per official site Supports uploading files or YouTuber links Per official description Per official description People who want to quickly process YouTube videos
Tinrec Free version offers basic quota for recording, import, summary, and transcription Focus on Chinese and mixed Chinese-English, multilingual meeting transcription Transcript, summary, chapters, to-dos, AI Q&A TXT, DOCX, SRT, VTT, etc., and export to third-party tools People who need meeting minutes, to-dos, and team knowledge retention

Stop organizing recordings by hand

Upload audio or video and automatically get a transcript, summary, and action items

Fields marked "per official description" mean I couldn't confirm from public information to a degree that would justify a definitive statement. I'd rather leave it blank than fill in a number that looks good.

User experience and suitable scenarios for each tool

Subanana: One of the few tools that treats Cantonese as the main language

Its design logic is clear: select Cantonese as the language, choose spoken or written output, and output in traditional or simplified Chinese. For people who need to deliver formal records, the time saved on "spoken to written conversion" is often more than the transcription itself.

The free part is "free preview of the first 15 minutes per file." Note that preview is not the same as full free export. If you need a full one-hour transcript, check the free scope before uploading, so you don't find out you have to pay after transcription.

Suitable for: interview transcripts, videos needing subtitle files, mixed Chinese-English content.

cSubtitle: An option when you don't want to register and only need subtitles

You can upload without an account, which is convenient for handling a single file on the fly. Upload capacity up to 4GB, duration up to 5 hours, supports MP4, MOV, MP3, WAV, M4A, OGG, WMA and other formats. Output includes text files, plus Word with time codes and .sub, .vtt subtitle files.

The downside is that language processing is Chinese-centric, with Cantonese as one option, so very colloquial usage still requires manual proofreading. The actual usable amount for the free version is described rather briefly; check the official site's current announcement.

Suitable for: those who only need subtitle files, handling a single file on the fly, and not wanting to create an account.

YouTuber AI: The easiest way to process YouTube videos

You can paste a YouTuber link directly to transcribe, which is convenient for people organizing video material—no need to download first and then upload. The official claim is up to 96% accuracy, but as mentioned earlier, such numbers only count when verified on your own audio.

Suitable for: organizing YouTube video content, taking notes, capturing key points from videos.

Tinrec: Transcription is just one part

Tinrec is positioned not as a simple transcription tool, but as meeting notes and collaboration. Besides uploading audio for transcription, you can use the desktop version to directly capture computer system audio to record online meetings on Zoom, Google Meet, Microsoft Teams, Webex, etc., without needing an extra meeting bot to join. After transcription, it generates summaries, chapters, to-dos, and allows AI Q&A around the meeting content, and exports to common formats and third-party tools.

To be honest: Tinrec's focus is on Chinese and mixed Chinese-English meeting scenarios. If you're recording pure Cantonese daily conversation, it's recommended to test a segment with your own audio first to confirm that names, sentence-final particles, and mixed Chinese-English recognition meet expectations before deciding on long-term use.

Also, for occasional recording processing, the free version is enough; only teams with weekly meetings need to consider the team version, which has independent team spaces, member roles and seat management, usage analytics, audit logs, etc.

Suitable for: online meeting notes, producing minutes and to-dos after meetings, and retaining meeting data as team assets.

Free quota and price comparison

Item Free content Paid plans (official announcements)
Subanana Free preview of first 15 minutes per file Per official site announcement
cSubtitle Free version, quota per official site Pro version, price per official site announcement
YouTuber AI Free tool Per official site announcement
Tinrec Free version basic quota Personal plan has weekly card, Pro monthly card, Pro annual card; team plan USD 29.80 per paid seat per month, USD 199 per paid seat per year

Two commonly asked comparisons: Notta free version is 120 minutes per month, max 3 minutes per single item; Otter.ai free version is 300 minutes per month, but mainly for English meetings and does not support Cantonese.

Currencies and promotions change; actual checkout amounts are subject to the official purchase page.

Conclusion: Choose based on your needs

Just want to quickly check if a Cantonese recording is worth organizing: Use Subanana's free 15-minute preview first, or paste a link directly into YouTuber AI.

Need to deliver subtitle files: cSubtitle can directly produce .vtt and .sub; Subanana also offers SRT and VTT.

Want meeting minutes, to-dos, and a searchable database: A tool like Tinrec that extends transcription to summaries, Q&A, and team spaces is more suitable, but verify Cantonese recognition yourself first.

Final reminder: The above is information verification and feature comparison, with samples limited to public information, and cannot be generalized to all recording scenarios. A quiet two-person interview and a noisy multi-person meeting will yield very different results. The most reliable approach is still to test with the most difficult recording you have on hand.

FAQ

Q1: Are free Cantonese speech-to-text tools really enough?

It depends on your use. If you just want to quickly grasp key points or confirm what the other person said, the free quota is usually enough. But if you need to deliver a formal transcript or subtitles, the free plan's "first N minutes" or "N minutes per month" can easily cut off midway, making things more troublesome.

Q2: Why do many tools claim "supports Chinese" but perform poorly on Cantonese?

Because when the language menu only has "Chinese," it's mostly a Mandarin model behind it. Cantonese has different vocabulary and word order, and forcing it will produce sentences where the characters are correct but the meaning is wrong. When choosing a tool, look for one that explicitly has "Cantonese" or "廣東話" in the language menu.

Q3: What's the biggest difference between free and paid versions?

Usually three things: available minutes, maximum file length, and advanced features (spoken-to-written conversion, speaker separation, subtitle format export). Some tools can transcribe in the free version but may not export the format you need—confirm this first.

Q4: How accurate is transcription for mixed Chinese-English Cantonese content?

This is where errors are most likely. It's recommended to prepare a recording containing company names, product names, and English terms for a real test, and check whether the tool supports custom vocabulary lists to add commonly used proper nouns.

Q5: Can the transcript be used directly as meeting minutes?

Not recommended to send directly. AI may miss words or misunderstand details. Important numbers, commitments, and names should still be checked against the recording timestamps. A safer approach: use the transcript as a draft, and separately confirm the summary and to-dos.

Q6: What should I pay attention to before recording?

First confirm the regulations in your location and company policies, and obtain consent from participants when necessary. In practice, placing the microphone closer to the speaker and reducing background noise will help recognition more than switching tools.

References

Turn every recording into actionable outcomes

Get 60 free transcription minutes when you sign in. No credit card required.

Upload audio or video for multilingual transcription, AI notes, and action items

Related Reading

You might also like

2026 Comparison of 4 Cantonese Speech-to-Text Apps: Which Saves More Time with Code-Switching and Post-Meeting Organization?

2026 Comparison of 4 Cantonese Speech-to-Text Apps: Which Saves More Time with Code-Switching and Post-Meeting Organization?

A one-hour Cantonese meeting often takes three hours to transcribe. This article reviews four tools that handle Cantonese speech-to-text, comparing them on mixed Chinese-English recognition, input sources, post-meeting organization, team data ownership, and free tiers, plus common pitfalls and scenario-based recommendations.

2026-11-17
4 Free Cantonese Speech-to-Text Tools Tested for 2026: Which Free Tier Is Enough?

4 Free Cantonese Speech-to-Text Tools Tested for 2026: Which Free Tier Is Enough?

I have over a dozen meetings a week, many with Hong Kong teams in Cantonese mixed with English. This article uses a mid-level manager's real work scenarios to compare the free tier algorithms, Cantonese recognition performance, and post-meeting organization capabilities of four Cantonese speech-to-text tools: Tinrec, Subanana, UniScribe, and cSubtitle. It also summarizes three common pitfalls to help you judge whether the free version is sufficient.

2026-11-17
How to Use iPhone Cantonese Live Translation in 2026: 5 Methods + 4 Key Settings

How to Use iPhone Cantonese Live Translation in 2026: 5 Methods + 4 Key Settings

A guide to 5 ways to achieve Cantonese live translation on iPhone in 2026, including Apple Intelligence translation in Messages, Phone, FaceTime, and AirPods, plus Tinrec's meeting transcription and translation workflow, with details on device requirements, language settings, and practical limitations.

2026-11-17
2026 Meeting Audio Transcription Buyer's Guide: 7 Solutions Compared and How to Choose

2026 Meeting Audio Transcription Buyer's Guide: 7 Solutions Compared and How to Choose

Meeting audio transcription isn't just about dropping a recording into a tool. This guide starts with the most common pitfalls, then reviews the positioning, quotas, and ideal users for 7 solutions—Tinrec, Otter, Notta, TurboScribe, Granola, NotebookLM, and CLOVA Note—and offers actionable selection advice to save you from re-listening to recordings after meetings.

2026-11-17
2026 iPhone Cantonese Live Translation Free Options: 3 Solutions Compared

2026 iPhone Cantonese Live Translation Free Options: 3 Solutions Compared

If you want to use Cantonese live translation on iPhone, there are actually 3 completely different paths: Apple's built-in Live Translation, the standalone Translate app, and Tinrec, which integrates translation into the meeting notes workflow. This article breaks down language support, use cases, and practical limitations item by item to help you determine which one truly solves your problem.

2026-11-17
iPhone Voice-to-Text Tested: Built-in Cantonese Transcription vs. Tinrec and Whisper in 2026

iPhone Voice-to-Text Tested: Built-in Cantonese Transcription vs. Tinrec and Whisper in 2026

iPhone's built-in Voice Memos can transcribe Cantonese recordings into text as of iOS 18.4, but if you need post-meeting summaries, action items, AI Q&A, cross-platform access, and team collaboration, tools like Tinrec can save you more time. This article compares iPhone's built-in feature, Tinrec, and offline Whisper solutions from a practical standpoint, with use cases, buying tips, and FAQ.

2026-11-17
2026 AI Real-Time Cantonese Translation Guide: No Bot Recording + Post-Meeting AI Summaries

2026 AI Real-Time Cantonese Translation Guide: No Bot Recording + Post-Meeting AI Summaries

A guide to choosing AI real-time Cantonese translation tools in 2026, comparing Tinrec, Subanana, OpenL Translate, and Transync AI, covering bot-free meeting recording, bilingual transcripts, post-meeting AI summaries, and team collaboration.

2026-11-17
3 Free iOS Cantonese Live Translation Options in 2026: Free Conditions and Use Cases

3 Free iOS Cantonese Live Translation Options in 2026: Free Conditions and Use Cases

A breakdown of the usable scope of iPhone's built-in Live Translation, AirPods Live Translation, and Tinrec's free version in Cantonese scenarios. Key clarification: although iOS 26.1 supports Cantonese for Apple Intelligence, live translation voice output may still be in Mandarin; the article compares free conditions, device requirements, operation entry points, and limitations.

2026-11-17
Free Cantonese Speech-to-Text Options in 2026: 5 Tools Compared

Free Cantonese Speech-to-Text Options in 2026: 5 Tools Compared

A student-budget guide to free Cantonese speech-to-text tools in 2026, comparing free tiers, Cantonese support, spoken-to-written conversion, and export features across Tinrec, Subanana, cSubtitle, Notta, and Otter.ai—so you can avoid the trap of choosing a free plan that isn't actually enough.

2026-11-17

Turn recordings into text, summaries and action items

Turn every conversation into clear notes and next steps with live transcription and AI summaries.

  • Audio to text
  • AI summaries
  • Automatic action items