How to Translate a Cisco Webex Meeting in Real Time — Chrome Extension Guide
Cisco Webex has its own answer to multilingual meetings, and on paper it looks complete. The Webex Assistant can transcribe a meeting and show translated captions in real time. The catch shows up the moment you actually try to turn it on: those features require an administrator-level license configured in Webex Control Hub. An individual user cannot enable real-time translation unilaterally — it is a tenant decision made by IT, tied to license SKUs, and rolled out at the organization level. And if you join a Webex call as an external guest from outside the host's organization, you do not get those features at all, no matter what the host org has licensed. Webex Assistant's translation language coverage is also narrower than what a dedicated translation engine offers, and Webex has no built-in way to speak back into a meeting in another language — there is no equivalent of a live AI interpreter voice.
This guide walks through a different approach: a Chrome extension that runs alongside the Webex web client and turns any meeting into a live, translated caption feed in 60 languages — without a Control Hub license, without an admin ticket, and without your meeting audio ever passing through a third-party cloud you did not choose. Total setup time is about two minutes.
What You Need
Before starting, make sure you have:
- A Chromium-based browser, version 114 or later. Chrome, Edge, Brave, Arc, and Opera all qualify. Safari and Firefox are not supported because they do not implement the extension APIs Gaavala depends on.
- A Webex meeting join link that opens in a browser tab — any standard
*.webex.commeeting URL works. - About two minutes to install the extension and start your first session.
You do not need a Webex paid plan, a Control Hub license, an administrator's approval, or any installation beyond the extension itself. The Webex desktop app is not used in this flow at all.
Step 1 — Install Gaavala from the Chrome Web Store
Open the Gaavala listing on the Chrome Web Store and click "Add to Chrome":
chromewebstore.google.com/detail/lfpocpgglghmaijacokoelmjlhipbgd
Chrome will show the standard permissions prompt listing what the extension can access. Approve it. Once installed, pin the Gaavala icon to your toolbar so you can reach the side panel quickly — click the puzzle-piece icon in the top right of Chrome and toggle the pin next to Gaavala.
When you click the toolbar icon for the first time, the Gaavala side panel opens and prompts you to sign in with Google or Microsoft. Sign-in runs entirely through OAuth using Chrome's identity flow — your password never reaches Gaavala, and there is no separate account to create on the Gaavala website. Your one-time free trial activates automatically and gives you 5 minutes of transcription total — no credit card required. Once you use those 5 minutes the trial is exhausted; it is a lifetime allowance, not a 7-day window and not a daily refill, so it never resets. The trial covers all 60 languages, speaker diarization, and transcript export as TXT, SRT, or JSON.
Step 2 — Join Your Webex Meeting in a Chrome Tab
Gaavala captures audio from a Chrome tab using the tabCapture API. That means the meeting must be running in a browser tab, not in the Webex desktop app.
Click your Webex meeting link. Webex will usually try to open its desktop application first. On the meeting landing page, look for the "Join from browser" option — Webex presents it as an alternative to launching the app. Click it, and Webex loads its web client in the current tab. If the page jumps straight into the desktop-app launch dialog, cancel that dialog; the "Join from browser" link typically appears on the landing page once the app handoff is dismissed.
Allow microphone and camera access if prompted, enter your name, and join the meeting normally. Leave this tab open for the duration of the meeting. You can switch to other tabs freely — the audio capture continues regardless of whether the Webex tab is in the foreground.
Step 3 — Open the Gaavala Side Panel and Press Start
With the Webex tab focused, click the Gaavala icon in your toolbar. The side panel opens on the right side of the browser. Click "Start" to begin a transcription session.
The moment you press Start, the extension captures the Webex tab's audio instantly through Chrome's tabCapture API — there is no screen-share dialog, no tab picker, and no checkbox to remember. Chrome shows a small indicator on the Webex tab confirming that it is being captured. This is normal and other meeting participants do not see it.
Behind the scenes, Gaavala routes the captured audio into a sandboxed offscreen document, opens a direct encrypted WebSocket (TLS 1.2+) to the Soniox speech engine in the EU, and starts streaming. The session uses a single-purpose temporary key that expires after 60 seconds. The audio is also piped back through the local audio graph, so you continue to hear the meeting normally — nothing is muted or redirected. No audio packet ever reaches a Gaavala server. If you want to verify that for yourself, the privacy architecture post walks through exactly how to inspect it in DevTools.
Step 4 — Pick Source and Target Languages
In the side panel, choose the source language (what the speakers are using) and the target language (what you want to read). Gaavala supports 60 languages through Soniox for both transcription and translation, including the long tail most tools quietly ignore — Turkish, Vietnamese, Tagalog, Czech, Greek, Hebrew, Hindi, Hungarian, Romanian, Thai, and more.
If your meeting has speakers in several languages, set source to "auto-detect". Soniox handles language switching mid-stream, and the captions update accordingly. Speaker diarization is on by default, so each utterance is labelled with a speaker tag — useful on a Webex call where the conversation moves quickly between participants.
You can change either language at any time without restarting the session.
Step 5 — Read Live Captions (or Speak Back with Pro)
Translated captions stream into the side panel within a second or two of each phrase being spoken. Click any segment to copy it, and export the full transcript as TXT, SRT, or JSON at the end of the meeting. If you are on Pro, you can also request an AI summary that condenses the conversation into bullet points and action items — that summary runs 100% on-device through Chrome's built-in model (Gemini Nano), with zero network calls, so the transcript text never leaves your machine.
If you upgrade to Pro ($24.99/month, billed through Lemon Squeezy as Merchant of Record, cancel anytime), you get 120 minutes of transcription per day — a daily quota that resets at 00:00 UTC — and the Speak Mode panel becomes available. Speak Mode is the two-way feature Webex has no equivalent for: you speak in your own language, and Gaavala synthesizes your translated voice into the meeting. There are two interaction modes:
- Push-to-Talk (PTT): hold a hotkey, speak, release. Gaavala translates and speaks the result into the meeting.
- Timed Speak: a hands-free 60-second window so you can deliver a prepared statement without holding a key.
While Speak Mode plays your synthesized voice, your real microphone auto-mutes so the two streams do not collide — and Webex's own mute button always wins, so you stay in control. Speak Mode offers three voice engines:
- Soniox studio voices — 28 multilingual presets that each speak all 60 languages, zero setup, the Pro default.
- Kokoro-82M — a compact model that runs entirely on your device via WebAssembly, free, with 25 English voices and no cloud round-trip.
- ElevenLabs (BYOK) — bring your own ElevenLabs API key to use voices you have already cloned in your own ElevenLabs account. Gaavala stores the key on your device and never creates voice clones itself.
Speak Mode is a Pro-only feature; it is not part of the free trial.
Common Issues
"Join from browser" link is not visible
Webex strongly prefers its desktop app and the browser-join option is sometimes one click below the fold. When you click the meeting link, cancel the desktop-app launch prompt — the "Join from browser" link usually appears on the landing page after the app handoff is dismissed. If you still do not see it, make sure you are on the meeting's web landing page (a *.webex.com URL) rather than a redirect that fires the app directly; reloading the page in the browser brings the web-join option back. In rare cases a host organization disables web-client join entirely at the Control Hub level, and there is no client-side workaround for that.
Captions show "no audio detected"
This is almost always a focus issue — the extension captures the tab that is active when you press Start. End the session, make sure the Webex tab is the focused tab, and click Start again. Other common causes: the Webex tab is muted at the OS level, the meeting has not yet started broadcasting audio, or you joined the audio via "Call me" / dial-in rather than computer audio, in which case the meeting sound is not in the browser tab for the extension to capture. Switch to "Use computer audio" in Webex and the captions resume.
Captions are accurate but feel lagged
Some latency is inherent — Soniox needs to hear enough of a phrase to commit to a translation. Typical end-to-end latency is one to two seconds. If it feels significantly longer, check your network: the audio streams over a WebSocket directly from your machine to Soniox, and a slow or lossy connection shows up directly in the caption delay.
Why the Extension Needs the Webex Web Client
A common follow-up question: why can the extension not just hook into the Webex desktop app? The honest answer is that browser extensions cannot reach desktop applications. Chrome's tabCapture API only sees what is happening inside the browser. A desktop Webex call runs in a separate native process with its own audio pipeline, completely outside Chrome's reach.
There are theoretical workarounds — virtual audio cables, system-level loopback drivers, native helper apps — but every one of them adds installation friction, signing requirements, and OS-specific bugs. The Chrome extension form factor was chosen specifically because it works the same on Windows, macOS, Linux, and ChromeOS, requires no driver install, and routes zero audio through any third-party server you did not pick. The trade-off is that you join meetings in the browser. The upside is that the exact same workflow — captions, hotkeys, language picker, Speak Mode — behaves identically whether the tab is running Webex, Teams, Zoom, or Google Meet. You learn one tool, not four.
What Gaavala Does That Webex Built-In Does Not
| Capability | Webex built-in | Gaavala |
|---|---|---|
| Real-time translated captions | Yes, but admin license required (Control Hub) | 60 languages, any pair, no license |
| Available to external guests | No | Yes — the extension is yours, not the host's |
| Language coverage | Narrower curated set | 60 languages (Soniox) |
| Speak back in another language (AI voice) | No equivalent | Yes (Pro) — PTT and Timed Speak |
| Works across Webex, Zoom, Teams, Meet | No (Webex only) | Yes |
| Audio leaves a vendor cloud | Yes (Cisco cloud) | No — direct to Soniox, never through Gaavala |
| Time to enable | Days to weeks (admin + license rollout) | About two minutes, you do it yourself |
The two rows that matter most in regulated and cross-org work are the last few. If your team uses Webex on Mondays, Teams on Wednesdays, and Zoom for client calls, you only learn one tool with Gaavala — and the audio path stays the same regardless of which platform is in the tab.
FAQ
How much does it cost, and do I need a Webex license?
Gaavala is independent of your Webex license entirely. The free trial is a one-time 5 minutes of transcription with no credit card, and it never resets. Pro is $24.99/month, billed through Lemon Squeezy (Merchant of Record), and gives you 120 minutes of transcription per day that resets at 00:00 UTC, plus Speak Mode. You can cancel anytime from the side panel. None of this touches Webex Control Hub or your organization's Cisco licensing.
Will other Webex participants know I am using Gaavala?
No. Gaavala captures audio passively from the browser tab and never injects anything into the meeting itself. There is no extra participant, no bot, no audio modification, and no notification to the host. Chrome's "Tab is being captured" indicator is visible only to you. The one exception is intentional: if you use Speak Mode on Pro, your synthesized translated voice is heard by the meeting — that is the whole point of speaking back.
Is my meeting audio uploaded to Gaavala servers?
No. Audio streams directly from your browser to the Soniox engine in the EU over an encrypted WebSocket (TLS 1.2+), using a single-purpose temporary key that expires after 60 seconds. Gaavala's backend is never in the audio path — it only handles sign-in and subscription state. This architecture is specifically why teams in regulated industries (healthcare, government, finance) can adopt it without triggering a full vendor-cloud compliance review for the audio stream. The full architecture is documented in the meeting audio privacy post.
Can a Webex external guest use this?
Yes — this is one of the strongest use cases. Webex Assistant's translated captions are gated to the host organization's licensed members, so external consultants, vendors, and partners joining a Webex call cannot access them at all. Because Gaavala is your own extension running in your own browser, it does not care whose tenant the meeting belongs to. As long as you can join the call in Chrome and hear the audio, Gaavala can transcribe and translate it.
Can I use Gaavala on a phone or tablet?
Not yet. Mobile Chromium browsers do not implement the extension APIs Gaavala depends on (tabCapture, offscreen, sidePanel). Use a desktop or laptop with a current Chromium browser.
Get Started
One-time free trial — 5 minutes of transcription, no credit card required, never resets. Speak Mode is Pro-only ($24.99/month, 120 minutes per day, cancel anytime). Sign in with Google or Microsoft after install, join your next Webex meeting from the browser, and you will have live translated captions in about two minutes — no Control Hub license and no admin ticket required.