Languages & Detection
Restrict recognition to the one to three languages spoken in a session with the language padlock, or let Kalima detect the language. Multilingual sessions, language badges, and all 60 languages.
Kalima recognizes speech in 60 languages. Pick the one to three languages that are spoken in a session and keep the padlock next to them closed: recognition then sticks to those languages, which gives the most accurate transcript. If you pick no language, Kalima detects it on its own. This page explains the padlock, how detection works, and how multilingual sessions appear in your transcript.
Pick your languages and close the padlock
Next to the languages you pick for a session sits a padlock. While it is closed, recognition is restricted to those languages, up to three of them. This is the largest accuracy gain available: no passage ends up in the wrong language, and strong accents are recognized far more reliably.
- Every new session starts with the padlock closed, using your saved default languages from Settings › Transcription › Default language hints.
- Three languages at most. If you pick a fourth language while the padlock is closed, Kalima asks which of the three it replaces.
- Live and uploads alike. The restriction applies to live recordings and to uploaded files.
- Fixed while recording. Languages and the padlock cannot change while a recording runs. Set them before you press record.
Lifting the restriction
If a session really can contain any language, click the padlock. Kalima asks you to confirm with Unlock anyway, because the transcript gets noticeably less accurate without the restriction. Often the better choice is to add the extra language instead, as long as you stay at three.
While the restriction is lifted, the padlock and the line beside it turn amber and the accuracy meter drops. Unlocking applies to that one session only: the next new session starts with the padlock closed again. To choose Any as the language, open the padlock first.
Add only languages that are really spoken. One language gives the best accuracy, and two or three keep it high. A language that isn't in the audio can pull recognition in the wrong direction.
Automatic language detection
With the padlock open, or with no language picked, Kalima listens to the audio and identifies the language being spoken. Any languages you did pick still act as hints that steer detection toward them.
This works the same way everywhere transcription happens:
- Live transcription: words appear in the detected language as you speak.
- Uploaded files: the recording's language is detected during processing and reflected in the finished transcript.
Multilingual and code-switching sessions
Plenty of real conversations slide between languages: a bilingual meeting, a language class, an international call where people drop into their native tongue and back. Kalima handles this in a single transcript. For a session with two or three known languages, pick all of them and keep the padlock closed.
When the spoken language changes mid-conversation, Kalima detects the switch and labels each part of the transcript with a small language badge showing which language that segment is in. A single session can contain several languages, each marked where it occurs, so a mixed-language record stays clear and readable.
Where language badges appear
Per-segment language badges show up directly in the transcript display in the Studio, next to the relevant lines. They make it easy to scan a multilingual transcript and see at a glance who spoke which language and where the conversation switched.
For bilingual sessions, pair your languages with Live Translation so everyone can follow along. One-way translation renders everything into a single language, while two-way translates between two languages in both directions.
Set languages for a session or an upload
To change the languages every new session starts with, open Settings › Transcription and edit Default language hints.
Every supported language
Kalima supports 60 languages. The same 60 are available for live transcription, for uploaded files, and as translation targets, and any one can be translated into any other. Here is the complete list, with the language code Kalima uses:
| Region | Languages |
|---|---|
| Western Europe | English (en), German (de), French (fr), Spanish (es), Italian (it), Portuguese (pt), Dutch (nl), Catalan (ca), Basque (eu), Galician (gl), Welsh (cy), Afrikaans (af) |
| Northern Europe | Swedish (sv), Norwegian (no), Danish (da), Finnish (fi), Estonian (et), Latvian (lv), Lithuanian (lt) |
| Central & Eastern Europe | Polish (pl), Czech (cs), Slovak (sk), Hungarian (hu), Romanian (ro), Bulgarian (bg), Slovenian (sl), Croatian (hr), Serbian (sr), Bosnian (bs), Albanian (sq), Macedonian (mk), Greek (el), Ukrainian (uk), Russian (ru), Belarusian (be) |
| Middle East, Caucasus & Central Asia | Arabic (ar), Hebrew (he), Turkish (tr), Persian (fa), Azerbaijani (az), Kazakh (kk) |
| Africa | Swahili (sw) |
| South Asia | Hindi (hi), Bengali (bn), Urdu (ur), Tamil (ta), Telugu (te), Marathi (mr), Gujarati (gu), Kannada (kn), Malayalam (ml), Punjabi (pa) |
| Southeast Asia | Indonesian (id), Malay (ms), Thai (th), Vietnamese (vi), Tagalog (tl) |
| East Asia | Chinese (zh), Japanese (ja), Korean (ko) |
The same list is searchable wherever you pick languages for a session or an upload.
Read-aloud (text to speech) covers all 60 as well, with one voice per language.
Tips for the most accurate detection
A few habits make a real difference to how reliably Kalima identifies and transcribes your languages:
- Pick the spoken languages and keep the padlock closed whenever you know what will be spoken.
- Use clean audio: a good microphone and the right audio quality mode help the engine hear distinct sounds clearly, which improves both detection and accuracy.
- Add session context for names, jargon, and proper nouns so specialized terms land correctly alongside the right language.
- Avoid very short clips when you can: a longer sample gives detection more to work with, especially for similar-sounding languages.
Detection confidence can dip on very short clips or noisy audio. If a language looks wrong, re-running with the correct language picked and the padlock closed (and cleaner audio) usually fixes it.
Frequently asked questions
No, but it is the most accurate way. Pick the one to three languages that are spoken and keep the padlock closed. If you pick none, Kalima detects the language on its own.
60 languages, covering the major world languages and many regional ones. Any of them can be detected automatically or picked for a session.
Yes. Kalima detects language switches within a single session and labels each part of the transcript with a language badge. For the best result, pick all the languages that are spoken, up to three, and keep the padlock closed.
Yes, that is its purpose: while it is closed, recognition is restricted to the languages you picked. With the padlock open, your languages only act as hints and Kalima still detects the language itself.
Up to three while the padlock is closed. If you pick a fourth, Kalima asks which one it replaces.
Every new session starts with the padlock closed, because restricted recognition is the more accurate setting. Opening it applies only to the session you opened it in.
This usually happens with very short clips, noisy audio, or two similar-sounding languages, or when the padlock was open. Pick the correct language, keep the padlock closed, improve your audio quality, and re-record or re-upload. Adding session context for names and terms can also help.