FR
For organisers ready the same day, nothing to deliver

Human and AI
remote simultaneous interpreting

No booth to build the day before, no receivers to count at the door, no deposit to advance. Your interpreters go on air from a browser; your attendees listen on the phone already in their pocket. One more language can be added mid-event, with no hardware and no technician. Bring your usual interpreters — or we book one for the date.

One free 4-hour event · no card required · nothing to install

Receiver Headset Booth LIVE General assembly Choose your language Channel 1 Channel 2 Channel 3 Floor No headsets to collect

What keeps you busy the day before

If you run multilingual events, you know this list by heart. SIWEL removes all of it.

The delivery slot

Booked three weeks ahead, confirmed twice, and it still turns up while the stage is being built.

Receivers at the door

Someone dedicated to handing them out, an ID held as deposit, and the hunt for missing units on the way out.

The booth in the room

Two square metres of floor, a slot in the seating plan, and a technician to bring back to dismantle it at night.

The language added at T-2

A speaker confirms late and one more language is needed: transmitter, channel, extra receivers, new quote.

Remote attendees

Whoever is not in the room hears no interpretation at all — or only through a second setup to install.

The deposit

Advanced before the event, returned after inventory, minus one mislaid headset.

Tell us about your event

S · the start

Your interpreters, or ours

They go on air from a browser, one per language. Don't have one for that date? Book an interpreter from the console.

i · the descent

No hardware

A link to share, a QR code at the door. Everyone listens on their own phone — no attendee app to install.

w · the waves

52 languages and dialects

Including Moroccan Darija, Algerian, Tunisian, Egyptian and Levantine Arabic — as they are actually spoken, not Modern Standard Arabic.

l · the finish

Options, if you need them

Live captions, AI interpretation, hybrid mode: switched on when you ask, never by default.

You have seen what SIWEL does. Here is what its name means.

SIWEL, letter by letter

Five letters, five promises. The logo stands up and reads top to bottom: follow it from the S down to the L.

S

Solution

A complete solution, not a component to integrate. You create the session, you share a link — that is all. No equipment to rent, no app to install, no IT project.

i

Intelligent

AI when you ask for it, never by default. Machine interpretation for an internal stand-up, human interpreters for a board meeting, and both in the same session if that is what you need.

w

Worldwide

Your audiences are not in the same place and do not speak the same language. SIWEL streams as many languages as you need, to listeners in the room and across three continents.

e

E-remote

Everything runs in a browser. Interpreters work from home or from their booth, attendees listen on their phones. No headsets to hand out, no console to install on site.

l

Liability

Who interpreted what, in which language, at what time: every session is traceable. Human or AI, the origin of each voice stays identifiable — which is what sensitive files require.

Automatic mode optional, never on by itself

52 languages, live voice and text

When no interpreter covers a language on the day, the automatic mode takes over: it listens to the speaker, translates, and speaks into the attendee's ear — who can read the text at the same time.

52

languages understood. The speaker talks, the system transcribes, the attendee hears their own language — with nothing added to the room.

7

ways of speaking Arabic. Modern Standard, Moroccan Darija, Algerian, Tunisian, Egyptian, Levantine, Jordanian. Most platforms file them under a single label.

Measured

one by one, not advertised. Every language was run against a reference corpus and carries its score in our code, not in a supplier's brochure.

  • Albanian
  • Arabic
  • Armenian
  • Azerbaijani
  • Belarusian
  • Bosnian
  • Bulgarian
  • Cantonese
  • Chinese (Mandarin)
  • Croatian
  • Czech
  • Danish
  • Dutch
  • English
  • French
  • German
  • Greek
  • Gujarati
  • Hebrew
  • Hindi
  • Hungarian
  • Indonesian
  • Italian
  • Japanese
  • Kazakh
  • Korean
  • Lingala
  • Macedonian
  • Malay
  • Moroccan Darija
  • Persian (Dari/Farsi)
  • Polish
  • Portuguese (Brazil)
  • Portuguese (Portugal)
  • Romanian
  • Russian
  • Serbian
  • Slovak
  • Slovenian
  • Spanish
  • Swedish
  • Tamil
  • Thai
  • Turkish
  • Ukrainian
  • Urdu
  • Vietnamese
  • spoken Algerian Arabic (Darja as actually spoken in Algeria — NOT Modern Standard Arabic)
  • spoken Egyptian Arabic (Masri as actually spoken in Egypt — NOT Modern Standard Arabic. Use إيه (not ماذا), ليه (not لماذا), دلوقتي (not الآن), عايز (not أريد), ح + verb for the future, بـ + verb for the present, مش (not ليس), كده (not هكذا). NEVER start a question with هل)
  • spoken Levantine Arabic (Syrian/Lebanese/Jordanian dialect as actually spoken — NOT Modern Standard Arabic. Use the everyday spoken forms: شو (not ماذا), ليش (not لماذا), هلق/هلأ (not الآن), رح (not سوف/سـ), كتير (not كثيراً), منيح/كويس (not جيد), مو/ما (not ليس), هيك (not هكذا), بدي (not أريد), عم + verb for the present. NEVER start a question with هل — ask it the way people speak. Write numbers and dates as spoken, e.g. خمستاعشر, تلاتين)
  • spoken South Levantine Arabic (Palestinian/Jordanian dialect — NOT Modern Standard Arabic)
  • spoken Tunisian Arabic (Derja as actually spoken in Tunisia — NOT Modern Standard Arabic)

A language missing from your programme? Ask for it — we measure it before answering, and we say no when the result does not hold up.

GDPR

GDPR compliant,
and processed in France.

Your event audio does not leave France. The service is hosted in France, transcription runs on our own machines in Paris, and so do the synthetic voices. That is every account's default setting, not an option you have to ask for — and we publish, on a separate page, exactly what is processed where.

France

for hosting and listening. Server in France, compute card in Paris, synthetic voices on our machines. The GDPR frames every step, and we name who does what.

15 min

and the text is erased. An automatic purge takes care of it: not a commercial promise, a server task that runs on its own.

0

recordings, unless you ask. And then it requires a declared purpose and legal basis before the first byte, then erases itself after 7 days.

Human interpretation produces no text at all: nothing to transcribe, nothing to keep, nothing to transfer. We name our sub-processors, we state what each one receives and when, and an account can be locked to "nothing leaves the server" — which is what sensitive files require, and we document it rather than simply asserting it.

Three minutes of setup, no logistics

Same day, from a browser. Nothing to rent, nothing to hand out.

1

You create the session

You set the floor language and the languages to interpret into. The listening link and its QR code are ready straight away.

2

Your audience scans

The QR code on a sign at the door, or the link sent the day before. Each person picks their language as they arrive.

3

Your interpreters go on air

One per language, several taking turns if the session is long. The console shows who is interpreting and how many people are listening.

What you stop renting

A quote for interpretation equipment hire means a booth, headsets, receivers, a technician on site and a deposit. SIWEL removes the whole line: your audience listens on the phone already in their pocket.

No equipment to rent, hand out or collect

No booth to build, no headsets to give out at the door, no rental receivers to count back at the end, no deposit. A QR code on a sign, or a link sent the day before.

Make an existing event multilingual

A room, a webinar, or a hybrid event running both at once. You change nothing about your venue or your video platform: SIWEL sits on top, the same day if you need it to.

Dialects, not just official languages

Someone from Morocco does not speak the Modern Standard Arabic of the evening news. Your interpreters hold Darija, Algerian, Tunisian and Levantine the way your audience actually speaks them.

Nothing is kept after the session

Neither the audio nor the captions. What remains is the usage line that justifies your invoice — and you can read it back whenever you want.

What you can add, if you need it

Human interpretation is enough for most sessions. Everything else is switched on account by account, at your request.

Option

Live captions and transcription

What is said appears as text alongside the voice. Useful in a noisy room, and for anyone who reads better than they hear. The text stays readable during the session and goes away with it.

Option

Book an interpreter

A form in the console: date, duration, floor language, target languages, location. The request reaches us and we call you back to confirm it.

Option

AI interpretation

The machine transcribes, translates and speaks. For a language no interpreter covers that day, or for an always-on helpdesk. In hybrid mode the interpreter holds the sensitive language and AI takes the rest.

A quote, not a grid

Your audience size, your languages, your duration: no two events are alike, and a price grid always ends up charging one client for what another one needs. Describe your programme and we price it on the real configuration — answer within one working day.

Request a quote

Public tender, purchase order or framework agreement: tell us the format your finance team needs. Write to us.

Try it on a real meeting

One free four-hour event, no card, with your own interpreters. Enough to run a full morning and judge it on the evidence rather than on a sales demo.