Convert Thai audio to text

By the Potok teamUpdated ~2 min read

You can convert Thai audio to text by sending the recording to the Potok bot on Telegram, with a caution. Thai is scored by character error rate, and published tests of others on large-v2, not our measurements, give 11.5. Expect more mistakes than in the recommended languages. Files up to 2 GB work.

Transcribe a recording

Send your voice message, audio or video to @potok_now_bot on Telegram and the text comes back in the chat.

Open @potok_now_bot in Telegram

Files up to 2 GB · export to TXT, MD, DOCX, PDF, SRT and VTT · speaker labels and timestamps · about 30 languages

Key facts

  • Thai has no spaces between words, so quality is reported as a character error rate, not a word error rate.
  • OpenAI names Thai as a language where the faster turbo variant gets worse, and a third-party turbo run gave about 13.4.
  • Tone marks and vowel signs sit above and below consonants, so one missing sign changes a word.
  • Spaces in Thai text mark phrases rather than words, so the line breaks may look uneven.
  • Thai names are long and have several romanised forms, so keep the original script of the name from the audio.

What a transcript looks like

Sample sentence in Thai
[00:00] Speaker 1: เรามาเริ่มประชุมกันเลย ก่อนอื่นเราจะทบทวนผลลัพธ์ของสัปดาห์ที่แล้ว จากนั้นจะวางแผนขั้นตอนต่อไป

Meaning in English: Let's start the meeting. First, we'll review last week's results, then plan the next steps.

Written by us to show the script and the layout of a transcript. It is not the output of the recognition engine.

Published quality

Quality for this language
LanguageVerdictPublished resultNotes
Thai (ไทย)Use with carecharacter error rate 11.5% (large-v2, FLEURS)
Whisper paper, Whale paper, Whisper turbo discussion
Published tests show a higher error rate for this language. Check a short sample first and plan to proofread.
Thai is written without spaces, so quality is reported as a character error rate. OpenAI itself names Thai as a language where the faster turbo variant gets worse, and a third-party turbo run gave about 13.4.

These numbers come from published tests by other people, not from tests by Potok. Your own audio, with its noise, accents and microphone, will differ.

How to do it

  1. Plan a full proofread. Treat the output as a draft and keep the recording close by.
  2. Open the bot. Open @potok_now_bot in Telegram and send the file.
  3. Use Upload for long ones. Up to 2 GB; the mini app's Upload button takes the biggest files.
  4. Check tone and vowel marks. Look for missing or doubled marks over consonants in a sample paragraph.
  5. Check names and numbers. Compare them with a written source.
  6. Export. DOCX or TXT to edit, SRT or VTT for subtitles.
What Potok can do
FeatureDetails
Where it worksTelegram bot @potok_now_bot; large files go through the bot's mini app
File sizeUp to 2 GB
ExportTXT, MD, DOCX, PDF, SRT, VTT
Speakers and timeSpeaker labels and timestamps can be switched on
LanguagesAbout 30 languages with good published results; a few more with warnings
TranslationThe transcript is not translated into another language
DictationA separate Potok desktop app lets you dictate text into any program; it recognises Russian speech

Common mistakes

Common mistakes and how to fix them
MistakeWhat happensFix
Publishing the draft unreadMistakes are more frequent than in the recommended languages.Proofread every line.
Looking for spaces between wordsThai does not use them that way.Read by phrase.
Expecting every tone mark to be rightA missing sign changes a word.Spot-check a sample.
Choosing a faster variant to save timePublished notes say Thai gets worse on turbo.Stay with the standard path.

More

See the full list of languages, the tools and the glossary.

Frequently asked questions

Why does Thai carry a caution?

Its published character error rate of 11.5 on large-v2 is higher than for the recommended languages.

Why a character rate?

Thai is written without spaces, so researchers count characters, not words.

Which model did the 11.5 come from?

The earlier large-v2 model, in published tests by other people. We did not measure it.

What is the size limit?

2 GB per file, with Upload for large ones.

Is the Thai translated?

No. The transcript stays in Thai, the same language as the recording you send, and it is never converted into another one.

Sources

Each source was opened by hand on the date shown. Pages change, so open the source if a detail matters to you.

  1. Word error rate (Wikipedia), Wikipedia · checked
  2. Whisper speech recognition system (Wikipedia), Wikipedia · checked
  3. Whisper paper, Table 13 (large-v2 on FLEURS) · checked
  4. Whale paper, Table 5 (Whisper large-v3 on FLEURS) · checked
  5. OpenAI Whisper discussion on the turbo model · checked

Try it on your own recording

Send your voice message, audio or video to @potok_now_bot on Telegram and the text comes back in the chat.

Open @potok_now_bot in Telegram