Free Transcription Plans in 2026: Yating (Taiwan) and 5 Tools Compared

The meeting ends, and the real trouble begins

“Who said that budget number last time?” The meeting ran 50 minutes, and your boss stumps you with one question.

You know someone said it, but you have to replay the entire recording to find it.

If the meeting also mixed in Taiwanese, or code-switched between Chinese and English, even “finding a tool that understands how we speak” becomes a research project.

Many people end up asking: which country made Yating Transcription?

On the surface it’s about nationality, but really it’s three practical concerns: where does my recording go, does the tool understand the way we speak, and is the free quota enough for me?

Let’s give the answer first, then talk about how to choose.

Which country developed Yating Transcription? The answer is Taiwan

Yating Transcription is developed by Taiwan AI Labs, an AI speech-to-text service built by a Taiwanese team.

The App Store page states it directly: 100% Made In Taiwan—funding, data collection, and developers all from Taiwan. It is also the first AI speech-to-text service developed specifically for Taiwanese accents and opened to the public.

Language support is its standout feature: Taiwanese Mandarin, Taiwanese, Cantonese, Japanese, and Chinese-English code-switching are all within its scope.

It is available as a web version (Yating Studio) and a mobile app. In the app, tap the recording icon at the bottom to record and transcribe in real time; when done, tap “Export” in the top right of the transcript screen to save the file.

Note: For details on free quotas and paid plans, refer to the official website’s current announcements. Also, user reviews on the App Store mention that with free use, accuracy is about 80%, so you still need some time to correct errors.

Simply put, it does a great job of turning Taiwanese accents into text.

But if you want to ask AI directly after a meeting “what were the conclusions and who owns the action items,” that’s a different level of tool.

Before choosing a transcription tool, understand these 4 key points

1. What language are you recording

Pure Chinese meetings are relatively simple; the hard part is mixed Taiwanese, Chinese-English code-switching, and accent differences. Rather than trusting “supports multiple languages” in the spec sheet, try a short recording of your own—you’ll know in ten minutes.

2. Do you just need a transcript, or a post-meeting deliverable

A transcript is raw material; summaries, chapters, action items, and meeting minutes are the finished product. If your job requires delivering something after the meeting, whether the tool can shorten that step affects your commute home more than transcription accuracy.

3. Where the data is stored and who can see it

Personal accounts and team spaces are different. Personal account data follows the person; team space data belongs to the team—this makes a big difference when members leave or hand off work.

4. How the free quota is calculated

“Free” usually has two limits: total minutes per month and maximum length per file. Some give a few hundred minutes per month; others only process a few minutes at a time. Check clearly before deciding whether to pay.

Tinrec — I’d suggest trying this one first

Tinrec is an AI meeting notes and collaboration tool for individuals and teams, available on desktop, web, and mobile. Its core is not just “speech to text” but turning meetings into data you can keep using.

During the meeting: no bot to invite

The desktop version captures system audio directly, so it works with Zoom, Google Meet, Microsoft Teams, and Webex without pulling an extra bot into the meeting room. For client meetings, this is practical—the screen stays clean, and you don’t have to explain “who this is.”

After the meeting: summaries, chapters, and action items in one go

After recording, it automatically generates summaries, chapters, and key points, and extracts action items from the discussion. Long meetings can be browsed by section, so you don’t have to scroll from start to finish.

Even better is AI Q&A. You can ask “who mentioned the budget last meeting” or “what were the client’s concerns about pricing”—it answers based on understanding the content, not by dumping a list of keywords. With export, results can go to Notion, Google Docs, OneNote, or Dropbox for further work, and you can generate reports, tables, and meeting minutes.

Cross-language meetings: original, translation, or bilingual view

During recording, you can switch between original, translation, or bilingual view. Chinese-English code-switching is a common scenario for it, and the team plan lets you maintain your own product names and technical terms with “team hotwords.”

Honestly, where are the limits

The free version gives a basic quota, suitable for trying it out; if you have several meetings every week, you’ll need to upgrade. Also, it does not promise a fixed accuracy rate—audio quality, accents, people talking over each other, and technical terms all affect results. For important content, cross-check with the recording timestamps.

One more thing to note: the team plan currently focuses on team-level sharing, with roles and seats managing governance and permissions. It is not a fine-grained system where you can set independent permissions for a single file.

Team use case

The team plan is a separate team space, not several people sharing one account. Roles of owner, admin, and member, plus seats control who can record, upload, edit, and export; members without a seat can still view, play, and read data.

Currently, a paid seat is USD 29.80/seat/month, or USD 199/seat/year (about USD 16.58/month). Each paid seat provides 2,000 minutes of team shared import quota per month; the first eligible team trial is 7 days, 1 free seat, 300 minutes shared quota. Real-time recording under an active plan is currently not deducted by minute; prices and promotions are subject to the official purchase page.

On the admin side, there are usage analytics, audit logs, and an audio recycle bin (30-day retention). When a member leaves, you can hand over their team resources to others—this is key to “data not following the person.”

Who it’s for: People whose meetings are mainly in Chinese, who need post-meeting summaries and action items, occasionally have cross-language meetings, or whose whole team wants to centralize meeting data.

Besides Tinrec, what other options are there?

Notta: Good multilingual transcription and translation, with team plans. Free version: 120 minutes per month, 3 minutes per item; Pro annual billing is about USD 8.17/month, 1,800 minutes per month. If your needs lean toward subtitles, translation, and batch file processing, it’s a reasonable option. But if you want post-meeting follow-up questions and team data retention for Chinese meetings, Tinrec’s workflow is more complete.

Otter.ai: A veteran for English business meetings. Free Basic: 300 minutes per month; Pro annual billing is about USD 8.49/user/month. If 90% of your meetings are in English, it will be smoother than Tinrec; but Chinese and Chinese-English code-switching are not its strong suit.

Granola: Also bot-free, focusing on “your handwritten notes plus AI enhancement.” Basic is free; Business is USD 14/user/month. Note that it does not support uploading pre-recorded audio files and does not save meeting audio—if you need archives for review, Tinrec keeps recordings and supports file import, which better fits that need.

Pitfall guide: 4 common mistakes when choosing a transcription tool

Pitfall 1: Deciding based only on “which country developed it”

Country does matter—it relates to language tuning and data flow. But it shouldn’t be the only criterion—what you really need to ask is “does it understand my meetings, and where is my data stored.”

Pitfall 2: Looking only at free, not at limits

The most common free-plan pitfall is per-file length. It looks like you get many minutes per month, but a 90-minute meeting can’t be uploaded, or you have to split it into several parts.

Pitfall 3: Treating the transcript as the end goal

A transcript is just raw material. If your tool only gives you a big block of text, the time to organize summaries and action items is still yours. Before choosing, think: how much of those two hours after the meeting does the tool save you?

Pitfall 4: Ignoring data ownership

If you use a personal account for meetings, the data stays in the personal account. If the team wants to accumulate meeting knowledge, use a solution with team spaces, seats, and handover mechanisms; otherwise, when a person leaves, the data goes with them.

Conclusion: Which one should you choose?

Back to the original question—Yating Transcription is developed by a Taiwanese team (Taiwan AI Labs), no doubt about it; it does have a local advantage in recognizing Taiwanese accents, Taiwanese, and Chinese-English code-switching.

But choosing a tool still depends on your workflow:

• Just want to turn Taiwanese-accented recordings into text, limited budget → try Yating Transcription first
• Mainly Chinese meetings, need post-meeting summaries and action items, and want to ask AI directly → Tinrec
• Meetings almost entirely in English → Otter.ai
• Batch transcription of many audio files, subtitle translation → tools like Notta
• Team wants to centralize meeting data, manage members and usage → Tinrec team plan

My usual advice is don’t rush to pay. Take a real meeting recording, run it through the free version, and see the error rate and whether the summary is usable—you’ll know in ten minutes.

References

• Yating Transcription
• Yating Transcription - App Store
• Yating Transcription
• What is Yating Transcription? 2026 Review of 7 Speech-to-Text Tools and Selection Guide - Tinrec Blog
• Generative AI Guide, National Kaohsiung University of Science and Technology - Teaching Resource Repository