Skip to main content
v2026.11,986 entries · CC-BY 4.0

Editorial · CASRAI · AI and ML research outputs

Otter adds 13 transcription languages and previews multilingual mode

Otter adds 13 transcription languages; 14 more and broader automatic language switching remain planned for a later rollout.

Published 16 Nov 2026· 4 minute read

Ask CASRAI · free to try

Ask about this story

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

CASRAI is the reference for research administration — bookmark it for the next question.

Otter.ai announced 13 new transcription languages on November 10, 2026, at Web Summit in Lisbon, while previewing 14 further languages and a broader multilingual mode. For research teams recording international collaboration or interviews, the immediate change is a wider language selection; automatic switching across the expanded set remains planned.

Which Otter languages are available now?

The announcement covers two batches: 17 languages are available, rising to 31 when the rollout finishes.

Release position Languages
Previously supported English, French, Spanish and Japanese.
13 additions available now Mandarin Chinese (Simplified), German, Italian, Cantonese, Korean, Portuguese (Brazil), Russian, Indonesian, Malay, Dutch, Turkish, Vietnamese and Polish.
14 announced for later Swedish, Danish, Finnish, Arabic, Hindi, Thai, Hebrew, Norwegian, Greek, Macedonian, Persian, Czech, Hungarian and Romanian.

TechCrunch reported the phased expansion. Its November 9, 11 p.m. PST timestamp falls on November 10 UTC.

Transcription, translation and language switching are different

Otter’s supported-language documentation says selecting a language produces text in the language spoken. Selecting German, for example, does not turn an English interview into German. AI Chat can separately translate conversation content, including summaries and action items, without replacing the original transcript.

For an interview project, that distinction separates the source-language record from an AI-generated translation used by collaborators. CASRAI’s explanation of transcription conventions for qualitative research provides background on how transcript choices affect the record researchers analyze.

There is also an existing exception to the broader switching preview: Otter’s language-setting guidance documents automatic French-English switching when French is selected. The newly announced multilingual option extends the ambition beyond that pair.

LeanStackAI’s independent analysis likewise emphasizes that the new languages are selected manually and the broader automatic switching feature has not shipped. Its analysis also distinguishes a transcript in the spoken language from translation into another language.

What changes for interview and collaboration records?

The language-setting documentation identifies practical differences across devices. Desktop and web imports provide a language selector; mobile imports use the account’s default conversation language. Selecting an import language on desktop or web updates the default for future conversations, too.

For a shared research workflow, the implication is that an import setting can affect the next recording. Changing the default does not automatically reprocess older conversations, and changing it during a meeting does not alter the language selected when recording began.

Only the conversation owner can reprocess an existing recording in another language, and that operation requires desktop or web. For teams sharing interview material, ownership therefore affects who can correct a language-selection error before colleagues begin coding interview data. Otter also cautions that AI feature availability and output quality can vary by language.

Accuracy needs similar care. Otter’s help page identifies accent, pronunciation, speaking speed, noise and recording quality as factors affecting transcription. The announcement’s benchmark claim concerns English; it does not demonstrate equivalent performance in the newly added languages.

Language coverage leaves data governance questions in place

Otter’s privacy policy, effective June 16, 2026, lists model improvement and training among its purposes for processing data. That includes de-identified audio recordings and transcripts that may contain personal information. It also says recordings can contain information about other people and asks users to obtain the necessary permissions before providing it.

The policy distinguishes ordinary service use from arrangements under a Business or enterprise agreement, where Otter says it processes information on the customer’s instructions and the customer acts as controller. For universities, the concrete implication is that language availability alone cannot establish which agreement and processing arrangement covers a research recording.

Otter’s separate privacy and security explanation says users must ask for consent and indicate that conversations are being recorded and transcribed. It describes user-controlled sharing permissions and says human review of audio requires explicit customer consent. Those statements concern different activities from automated model training; consent to one should not be read as a description of the other.

The same explanation says Otter’s external AI service providers do not train their models on customer data. That assurance concerns those providers, while the privacy policy separately describes training Otter’s own technology. This distinction matters when a university interprets a vendor’s data-use assurances.

Deletion also has a defined workflow: the company says deleted conversations move to Trash, where automatic deletion follows after 30 days unless the user clears them sooner. That detail makes recording access and deletion relevant to a project’s data management plan, alongside the transcript language and any translated derivatives.

What happens next?

Otter has announced no release date for the remaining languages or the broader switching mode. Its help center invites interest in testing upcoming languages, while stating that submitting the survey does not guarantee beta access.

The company is soliciting requests for additional languages as well, but says a request does not establish a release timeline. For collaborations needing a language in the later batch, the announcement therefore provides a roadmap rather than a dated commitment.

Sources

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →