Frequently asked questions

sotonis takes the microphone audio in your venue and puts real-time interpretation captions on the stage screen and on every guest's phone (scan a QR — no app). It is not a personal translation app or a video-call captioner; it is built for in-person events like church services, seminars, and conferences. Pick any two of 82 languages and it works both ways.

sotonis also puts captions on the stage screen — something phone-only apps and video-call captions can't do. Instead of everyone staring at their own phone, one screen up front reaches the back row and guests who don't use phones. Just connect a mic; guests only scan a QR.

No. Scanning the QR opens the captions right in the phone browser — no download, no sign-up. Each guest can pick text size and whether to see the translation, the original, or both.

The free plan gives you 3 hours a month, up to 30 minutes a day, one-way. One free account per computer and per network (IP) — extra accounts will not open on the same computer. Per-event: $59 for 1 hour, $99 for 2 hours, $250 for a half-day (6h), $400 for 24 hours. For weekly use, subscriptions are cheaper — Starter $47/mo (6h), Basic $119/mo (12h), Plus $199/mo (25h), Premium $279/mo (39h), Enterprise $359/mo (55h), on an annual contract (launch prices, 20% off). Starter runs $15/hour beyond 6 hours. Churches and nonprofits get an extra 20% off on a 12-month plan. There is no charge based on audience size.

No. All 82 languages are included, and adding a language doesn't raise the price. There's also no per-attendee fee.

Yes — churches and nonprofits get 20% off. It applies to weekly-use subscriptions, and the annual contract makes budgeting easy.

Much cheaper. Human interpreters typically run hundreds to over a thousand dollars a day, per language direction. A sotonis all-day pass is $400, and it costs the same no matter how many people attend.

82 languages, and you pick two to run both ways. For example, a pastor preaches in Korean and English captions appear; an English speaker talks and Korean captions come back — in real time.

Yes, on Plus and above. Guests who find captions hard can plug earphones into their own phone and hear the interpretation as audio (Starter and Basic are captions only). Captions remain the core, though — no receivers to hand out, collect, or charge, and everyone follows one stage screen.

Yes — it can caption audio playing on your laptop, such as a video-call speaker output or a screening. Those captions go to the stage screen and to the phones of people in that room. It is not a video-call integration that distributes captions to each remote participant, and sotonis is designed primarily for in-person events.

Yes. Upload materials like scripts, the program, or speaker bios, and the AI extracts names, organizations, and terms; after you review them, those exact spellings are applied to the captions. Names are what AI interpretation gets wrong most often, so this step moves quality more than anything else. We test the registered spellings against real pronunciation at rehearsal, so event day starts from a verified list.

It detects the direction automatically and captions the other side. Real events are rarely monolingual — the opening in one language, a greeting in another, a panel that mixes both. sotonis identifies the language being spoken and captions its counterpart, following the speaker even mid-sentence. The stage screen shows source and translation together, so neither side of the room is left out.

No limit. Even 1,000+ read on their own phones at no extra cost, with no devices to hand out.

No — approved texts are never re-translated by the AI. The Lord's Prayer, the Creed, and the benediction appear word-for-word in Worship Mode, and Scripture follows the wording your church registers, so the AI never invents the text.

Worship Mode speaks the church's language. Terms like justification, sanctification, and intercession stay consistent throughout, and denomination-specific wording is registered ahead of time — not re-guessed each time like a generic translator.

That's exactly why we built it. The pastor preaches in Korean and English captions appear on the screen and their phones; English testimony comes back in Korean for the elders. One service, no generation left behind by language.

It runs Google Gemini 3.5 through a real-time API — a translation-native model, so the translation keeps up before a sentence ends. Pre-registered terms and a rehearsal push event-specific accuracy higher. Like a human interpreter, it listens then speaks, so there's a natural few-second delay.

A separate PC for the interpretation, your event mic (mixer or USB), and an HDMI screen. Keep it apart from the PC running your slides. No IT background needed — we set it up together at a pre-event rehearsal, and there's a 5-minute how-to video.

We recommend a separate PC. On one machine, the slideshow and the caption output window compete for the same display, and a heavy presentation file pushes interpretation behind. It doesn't need to be powerful — any laptop with an internet connection is enough.

About 0.5 Mbps up and 1 Mbps down is enough. Data usage is roughly 0.7 GB per hour — about 2 GB for a 3-hour event. Anywhere with a normal LTE signal, a phone hotspot can run the whole thing (for weak-signal venues we recommend a pre-event check). Audience phones receive text only, so they use almost no data.

Audience screens reconnect automatically and hold the last caption. While the operator PC is offline, though, no new captions are produced — interpretation resumes on its own once the line returns. That's why we recommend a dual setup as standard: wired internet plus a phone hotspot. The bandwidth need is small enough that one hotspot covers the backup, and we test the switchover with you at rehearsal.

Yes. If phones get a signal, one LTE hotspot is enough — a 3-hour event uses about 2 GB, well within a normal data plan. Where there's no cellular at all, a portable satellite terminal (Starlink Mini and similar) covers it, and quality is identical either way. We don't offer a fully offline mode: on-device interpretation is noticeably worse today and didn't pass our bar. If your security policy forbids outbound audio, tell us and we'll factor it into the roadmap.

Audio is streamed to the interpretation engine (Google Gemini API) solely to generate captions. sotonis servers never store audio and no recording is made, and under the paid API terms the input is not used to train AI. A printable data-handling notice for your church board is at sotonis.com/data.

If captions don't show on event day, you get a full refund. Single events are fully refundable any time before you start interpretation, with no contract. Subscriptions have a 3-month minimum and are fully refundable within 7 days if unused. Full details at sotonis.com/refund.