Protocol locked in the glossary
Names, titles, forms of address and agreed designations for countries and territories pinned exactly, so a translation cannot improvise on a sensitive name.
For diplomacy & international organizations
Embassies, foreign ministries and multilateral bodies now publish more video than their translation services were staffed for: ambassador messages, public diplomacy campaigns, side-event recordings, explainers for member states. Octavia localizes that surrounding material with names and titles locked, and leaves the official record to the services built for it.
The specific reasons this content usually stays in one language.
Official texts in the official languages are rightly prioritised; video, social content and recordings queue behind them, often indefinitely.
A minister’s title rendered informally, a country named in a disputed form or an honorific dropped is read as a position, not a typo.
Public diplomacy reaches people in their own language, including regional and minority languages a mission’s small team cannot cover.
The parts of the platform that matter for this work, rather than the full feature list.
Names, titles, forms of address and agreed designations for countries and territories pinned exactly, so a translation cannot improvise on a sensitive name.
Diarization keeps each delegate and panellist distinct across a long recording, so subtitles and transcripts keep every statement with its speaker.
Ambassador messages, cultural programmes and consular explainers localized for host publics — subtitled for social feeds, dubbed where audiences expect to listen.
60+ languages, so member-state and host-country audiences outside an organisation’s official languages can be reached without a vendor per language.
Transcript and translation go through your own approval chain — desk officer, press office, native-speaker reviewer — before any audio is generated.
Role-based access with SSO, encryption in transit and at rest, configurable retention and no training on your content, for statements prepared ahead of an embargo lifting.
Four steps from the file you already have to every language you need.
Official texts stay with the translation service; Octavia takes the video and outreach around them.
Names, titles, honorifics and agreed designations in each target language, approved once by protocol.
Desk, press office and a native-speaker reviewer sign off the text before any audio exists.
Subtitled and dubbed versions for host publics, member states and the archive.
One upload in, every format you need to publish out.
Including the ones where the honest answer is a limitation.
No, and it should not try. Official texts in an organisation’s official languages — resolutions, communiqués, agreed conclusions — carry negotiated terminology and are produced and verified by the organisation’s own translation service. Octavia is for the material around that record: video statements, outreach, recordings and explainers. Where a video quotes an official text, use the published official translation rather than a fresh one.
Put it in the locked glossary before anything is translated: names, titles, forms of address and your agreed designations for countries and territories, in each target language. Locked entries are used exactly as approved rather than retranslated. What a glossary cannot catch is tone — the register of a condolence message, the weight of a qualifier — so a native-speaker reviewer who knows the context should still approve every language.
Do not plan on it. Interpretation in the room is a professional discipline and, in formal meetings, part of the proceedings. Octavia’s place is afterwards: translated subtitles and transcripts of the recording for delegations, the public and the archive, and dubbed statements for audiences who would otherwise never watch them. Mark translated meeting transcripts as unofficial so they are never mistaken for the verbatim record.
Only with their explicit consent, on statements they have approved word for word in each language, and with the synthetic voice disclosed. A cloned voice delivering words a principal never cleared is an incident waiting to happen. Many missions use a professional synthetic voice per language and subtitle principals instead. Non-public or sensitive material should stay on systems your security office has approved for it.
Practical guides from the Octavia editorial team.
The same engine, pointed at a different kind of work.
Start with one recorded statement, subtitled in the host country’s languages, with protocol names locked first.