AI interpretation · Live captions · Multilingual events

Live translated captionsthat keep every audience in the conversation

Start with a clean audio feed. WALI connects live transcription, AI translation, approved terminology and operator correction, then delivers the same captions to venue screens, OBS production and audience phones.

Project consultationComplete a one-minute briefSent directly to zoom@fanyi.im
  • Interpretation Caption SaaS
  • Industry terminology
  • Translation quality control
  • Live operator correction
INTERPRETING LIVE VENUE MAIN
EN
SOURCE CAPTION

Please refer to the bilingual captions on the main screen.

ZH
TRANSLATED CAPTION

请关注会场主屏上的双语字幕。

01 INPUTLive audio 02 ASRRecognition 03 TRANSTranslation 04 OUTPUTDelivery
PROVEN IN PRODUCTION A live-caption workflow tested in real venues.

Every capability described here is backed by shipping software and traceable project records.

18traceable project records
7languages in one event workflow
Onsite + abroaddomestic and cross-border delivery
Multi-surfacescreens, production, web and mobile

CAPTION OPERATIONS SAAS
INTERPRETATION WORKSPACE

Prepare the room once, then bring the same settings into production.

Rooms, languages, provider routes, terminology, caption presentation, accounts and usage are managed in one place. Prepare projects in the browser workspace, then run the live event from the Web or Windows client.

Caption operationsINTERPRETATION WORKSPACE
Workspace connected
01Rooms 02Languages 03Terms 04Delivery
Enterprise workspaceInterpretation roomsOne configuration
LOCAL
Local caption roomWindows processes the direct route while the platform synchronises caption state and outputs
Local route
ONLINE
Online caption roomThe platform processes audio, recognition, translation and realtime delivery
Platform route
ConfigureRecognise & translateCorrectDeliver
OperatorCorrectionVenue screenOBSAudience QRVoice output
01 / ROOM

Central room and route configuration

Save source and target languages, recognition and translation routes, caption pacing and delivery settings instead of rebuilding every event.

02 / LANGUAGE

Project terminology across the language chain

Keep speakers, brands, products and domain terms in one project while preserving the distinct roles of recognition terms, approved translations and replacements.

03 / DELIVERY

One caption state for every output

Operator, correction, clean projection, OBS, audience QR and voice output share the same room state instead of maintaining separate copies.

04 / TEAM

Enterprise accounts and auditable usage

Enterprise main accounts can manage subaccounts and available time, while billing records show how platform functions are consumed.

Run the topology assigned to the accountLocal-direct and online-platform rooms share one management and delivery experience while retaining different audio-processing routes.

Open Caption SaaS Download Windows client

TRANSLATION QUALITY
TERMINOLOGY CONTROL

Accuracy is not a slogan.It is a workflow that can be prepared, monitored and corrected.

Live translation is affected by audio, accent, context and specialist language. WALI brings industry instructions, recognition terms, approved translations, deterministic replacements and human review into one caption workflow to reduce predictable, manageable errors.

Terminology policy active 中文 ⇄ EN
INDUSTRY TEMPLATE Artificial intelligence & machine learning Context instruction + bidirectional terminology
智能体AI agentmatched
检索增强生成retrieval-augmented generationmatched
推理时计算inference-time computematched
27industry and specialist templates 5,000+Chinese-English term pairs Two-waymatching by translation direction
01 / SOURCE

Start with a reliable source transcript

Check levels, noise and source language, then apply names, brands, models and abbreviations as recognition terms before translation begins.

02 / CONTEXT

Give the model the right domain context

Select a concise industry instruction for manufacturing, finance, medicine, AI, energy and other domains, preserving figures, units and non-expansion boundaries.

03 / TERMINOLOGY

Keep critical terms in one approved form

Only terms matched in the current speech are added to translation context, then applied for the actual target language to reduce drift during long events.

04 / REVIEW

Keep critical facts open to correction

Operators can correct names, amounts, dates and key statements in real time, while deterministic errors are handled by explicit replacement rules.

HOW IT WORKS

Audio comes in. Clear captions reach every screen that needs them.

Audio ingest, live recognition, machine translation, operator correction and caption delivery are managed in one workflow for consistent monitoring and control.

STEP 01

Audio ingest and quality check

Connect the mixing console, conference microphones, computer audio or a remote meeting feed, then verify levels, noise and network readiness.

STEP 02

Live transcription and terminology

Generate source captions continuously while applying prepared speaker names, brands and specialist terminology for the event.

STEP 03

Translation and operator correction

Produce translated captions for the configured language direction, with live monitoring and correction for critical information.

STEP 04

Synchronized multi-channel delivery

Publish one managed caption stream to venue displays, production systems, web viewers and audience devices.

One workflow for recognition, translation and deliveryLess tool switching for the production team, consistent content for every viewer.

CORE CAPABILITIES

Live production needs more than a caption overlay.

Terminology, caption presentation, operator controls and delivery channels work together to support accurate presentation and reliable event delivery.

MULTI-SCREEN DELIVERY

One caption source across every output

Source and translated captions are published from a single managed stream to venue, broadcast and audience endpoints.

LIVE CAPTIONS

Continuous captions for live speech

Designed for keynotes, panel discussions, corporate training and audience Q&A, with source and translated text delivered in real time.

PROJECT TERMINOLOGY

Event-specific terminology

Prepare company names, products, speakers and domain terminology from agendas, scripts and presentation materials.

LIVE CONTROL

Centralized recognition, translation and output control

Operators can monitor service status, correct critical text, configure source and translation styles, and manage separate output channels.

  • Operator correction for key captions
  • Separate source and translation styles
  • Independent venue and audience outputs

REAL-TIME MULTILINGUAL CALLS

Multilingual audio and video calls that keep global work moving.

Enterprise users start an audio or video call from Android and select the language pair required by the conversation. Invited guests join from Android or a browser, while one consistent source-and-translation caption stream is shared across the room.

Configurable language pairs One shared caption stream Browser guest access
01

Language routing for each project

Select source and target languages for the business context. Chinese-English supports continuous automatic direction; other combinations follow the selected language pair and available speech capability.

02

Shared source and translation for every participant

The room maintains one authoritative caption feed, keeping hosts and invited participants aligned without repeated explanations.

03

Browser access with no pre-installation

External guests open an invitation link, grant microphone or camera permission, and join from a browser when appropriate.

04

Separate handling for captions and media

Captions may be retained for business needs, while audio and video are used for live transmission and are not stored as complete recordings by default.

Discuss an enterprise call workflow
REAL PRODUCT UI Browser meeting view
meeting.fanyi.im / multilingual-call
Four business participants in a WALI real-time multilingual audio and video call 中文 English · Captions connected
One shared caption stream
我们今天先确认产品演示和交付时间。Today, we’ll confirm the product demo and delivery schedule.

PRODUCTS & SERVICES

One enterprise language workflow for live, remote and content operations.

Select software capabilities or managed delivery according to whether the content is live, how audio enters the system, where captions must appear and what level of operational support is required.

AI Cloud Notes

Capture meeting audio on Windows, identify speakers in real time and develop the transcript into an editable, searchable and exportable formal record.

  • Speaker diarization
  • Live correction
  • Minutes and DOCX

Caption SaaS and managed event service

Manage rooms, languages, terminology, correction and delivery in one workspace, with rehearsal and onsite support available for managed projects.

  • Room management
  • Translated captions
  • Multi-channel delivery

Broadcast and production captions

A clean, control-free caption source for production systems or streaming software to place source and translated text in the program frame.

  • Clean caption source
  • Visual configuration
  • Production integration

Audience captions and translated audio

Audience members use a QR code or link to read live captions—and, when configured, listen to AI-generated translated audio.

  • No-install access
  • Live captions
  • AI voice output

Multilingual audio & video calls

Select a language pair, start from Android, invite Android or browser guests, and share the same live source and translated captions.

  • Multilingual translation
  • Shared captions
  • Browser guests

Dual-screen translation terminals

Independent displays for each speaker in reception, exhibitions, hospitality desks and service counters.

  • Dual displays
  • Two-way voice translation
  • Scenario customization

Text, image and offline translation

Translate typed content and images, with offline recognition and translation available on compatible devices for mobile and limited-connectivity use.

  • Text translation
  • Image translation
  • Offline capability

AI Voice Studio and SSML

Direct voice, emotion, pace and pauses by scene and paragraph, then retain every generated version and download approved audio.

  • AI voice direction
  • SSML control
  • Preview version history

Media transcription and subtitling

Transcribe, translate, review and format existing audio or video, with multilingual subtitle files and captioned video available for publishing.

  • Media transcription
  • Subtitle files
  • Captioned video

Enterprise customization and integration

Configure branded interfaces, language workflows, caption endpoints, account permissions and integration with existing business systems.

  • Brand and UI
  • Endpoint fit
  • API and workflow integration

TRUST BY DESIGN

Professional delivery begins with preparation and live control.

Meeting language services depend on content, production audio, networks, display systems, broadcast workflows and onsite personnel. Each managed project defines these responsibilities and validates the critical route before the event.

01 / PREPARATION

Pre-event materials and terminology

Use agendas, slides, scripts and speaker information to prepare recognition terms, translation terminology and replacement rules.

Create a language asset specific to the event
02 / ACCESS

Audience access and endpoint planning

Define caption layouts and access for venue displays, broadcast frames, web viewers and audience devices.

Match the output to the way people actually attend
03 / OPERATION

Live monitoring and correction

Operators monitor recognition, translation and output status, correcting speaker names, figures and critical statements where required.

Combine automation efficiency with human control
04 / REHEARSAL

End-to-end rehearsal and fallback

Validate production audio, networks, displays, broadcast output and audience access using the actual event route.

Verify every critical interface before the event
Delivery principles

Capability boundaries, data handling and support scope are agreed before delivery.

  • Recognition and translation are evaluated against the actual audio, languages and specialist content
  • Project scope, language directions, equipment and support model are agreed in advance
  • Call media is live by default, not retained as a complete recording
  • Caption data and enterprise terminology are managed under permissions and project terms

USE CASES

Configure the solution around the meeting format, audio route and viewing endpoints.

In-person meetings, broadcasts, hybrid events and remote calls have different requirements for audio, caption delivery and operational support.

SCENE 01

International conferences

Live transcription and translated captions for keynotes, panels and Q&A, delivered to the main screen and audience viewers.

SCENE 02

Broadcast events and product launches

A clean caption source enters the production workflow while venue and mobile endpoints remain synchronized.

SCENE 03

Corporate training

Onsite and remote learners follow the same content, supported by product and course terminology prepared in advance.

SCENE 04

Cross-border meetings and technical support

Shared source and translated captions across configured language pairs for product demonstrations, project coordination, technical support and distributed teamwork.

SCENE 05

Reception and service counters

Use continuous meeting captions, dual-screen terminals or face-to-face translation according to the interaction.

FIELD RECORDS
PROJECT CASES

Different venues bring different constraints—and different answers.

Our records cover multilingual conferences, global product launches, industry forums and overseas business events, with emphasis on language configuration, output endpoints and delivery method.

PROJECT SUPPORT

Think through the details before show day.

Production events connect audio, networks, displays, broadcast systems, audience access and onsite personnel. We assess the requirements, configure terminology and caption presentation, validate the equipment route and provide the agreed level of operational support.

Caption visual designBrand, color, type size, layout and background
Terminology preparationCompanies, products, speakers and industry terms
Signal and output setupAudio, display, production, web and mobile viewers
Training and operationOperator training, technical rehearsal and event support within the agreed scope
01

Requirements and solution assessment

Confirm the event format, language directions, scale, audio source, caption endpoints, network and support model.

02

Terminology and visual configuration

Prepare event terminology and define the source-and-translation layout, branding, type size and display treatment.

03

Production-route rehearsal

Validate audio, recognition and translation, venue displays, broadcast output, audience devices and agreed fallback paths.

04

Live operation and project review

Provide remote or onsite technical support as agreed, then organize caption files and operational records after the event.

FAQ

Key questions before procurement and implementation.

What does the interpretation caption SaaS manage?

Enterprise and event teams can create and manage caption rooms, configure languages, recognition and translation routes, project terminology and caption presentation, then deliver one room to the operator, correction view, venue projection, clean OBS source and audience QR page. Each room uses the local-direct or online-platform topology assigned to the account.

How do live interpretation captions differ from post-produced subtitles?

Post-produced subtitles are created after recording. Live interpretation captions process speech as it happens and must account for production audio, recognition and translation latency, terminology, operator correction and delivery to venue, broadcast and audience endpoints.

How are captions integrated with venue displays and broadcast production?

Venue displays can use a dedicated projection view, while production teams receive a clean output without operator controls. Source-and-translation layout, type size, colour, background and brand elements can be configured for the stage and production frame.

Does the audience need to install an application?

Usually not. Audience members can scan a QR code or open a project link in a mobile or desktop browser. When AI voice output is included, it can also be delivered through the configured audience experience.

How are speaker names, brands and specialist terms handled?

Agendas, slides, scripts and speaker information can be used to prepare a project lexicon. Recognition terms improve source transcription, translation terms constrain the target form, and replacement rules handle deterministic spelling. WALI also provides 27 industry and specialist templates, which can be combined with event material and human review.

How does a multilingual audio or video call work?

An enterprise user starts the call from Android and selects a source and target language. Participants join through Android or a browser invite and share one source-and-translation stream. Chinese-English supports continuous automatic language direction; other combinations and auto-detection behaviour follow the selected language pair and available speech capability.

What affects recognition and translation quality?

Results depend on audio quality, accents, speaking speed, venue noise, languages and specialist content. Managed projects reduce risk through audio checks, terminology preparation, replacement rules, operator correction and rehearsal. Legal, medical, financial and other high-risk content may require human review or professional interpreters.

What should be prepared before a production event?

Provide the agenda, speaker list, slides or scripts, language directions, venue plan and technical information. Confirm mixing-console audio, network access, display endpoints, broadcast integration and operators, then rehearse the complete route with production equipment.

How are call media, captions and project data handled?

Call media is used for live transmission and is not retained as a complete long-term recording by default. Captions, settlement and necessary audit information are handled according to business and compliance needs. Access to terminology, captions and project material is controlled under the applicable terms, privacy policy and project agreement.

When should we use the software directly or request managed delivery?

Routine internal meetings or teams with their own operators can begin with the client software. Large venues, broadcast integration, multilingual launches, branded visuals, specialist terminology or onsite support should be assessed as a managed project.

PROJECT INQUIRY
Talk to the project team

Tell us what the venue looks like. We will help shape the right setup.

Include the event type, date and location, language directions, audience size, production audio and caption endpoints. The project team will assess software configuration, technical delivery, equipment integration and customization scope.

How is the inquiry handled?It is delivered to zoom@fanyi.im and used only to assess and respond to this project.
What information is useful?Date, location, language directions, audience size, production audio, broadcast needs and caption endpoints.
Still planning the project?Submit the known requirements first; the project team can confirm the remaining technical parameters by email.
Or email zoom@fanyi.im directly
01 / BASIC INFORMATION

Project requirement details

Fields marked * are required. Information is used only for project assessment and business communication.

The project team will respond by email after submission.

PLAN THE COMPLETE LANGUAGE WORKFLOW

Plan the right language service
for your next meeting or event.

From event format, language directions and audience scale to audio integration, caption endpoints and technical support, the project team will recommend an implementation based on the actual production environment.

Caption SaaSRequirementsTerminologyVisual setupRehearsalLive operation