Live Translate FAQ
Answers to common setup and operating questions about Live Translate.
For what the app is and how to set it up from scratch, see the
Live Translate Overview.
Setup
Do I need to configure a translation provider before my team can use it?
No. If you connect your own provider (AWS Translate or DeepL) at Admin → Organization → Integrations, Live Translate uses it. Otherwise it falls back to a platform-managed engine automatically. Speech-to-text runs on the platform’s OpenAI connection, so it must be available for a turn to run; if text-to-speech fails, the turn still arrives as on-screen text.
How do I make Live Translate visible to employees after I enable it?
Enabling the app at Admin → Apps → All Apps makes it visible to admins immediately. For employees and managers to see it, you also need to publish it to users — either globally or by setting up visibility rules for specific groups. Until you do, only admins and people you’ve added as the app’s App admins or contributors can open it. An App admin or contributor can use Live Translate before it’s published, but company terminology and flagged translations stay with company admins.
Permissions and access
Who can manage the glossary and view translation analytics?
Only admins and super admins. The glossary and analytics live at Admin → Translation, which requires admin-level access. Employees and managers can use the interpreter and flag translations as wrong, but they cannot add, edit, or delete glossary terms or view the analytics dashboard.
Why can’t some of my employees see Live Translate?
Three things must all be true: the app is enabled at Admin → Apps → All Apps, it is published to users (or the employee has an explicit visibility grant), and the employee’s account is active in the organization. If all three are in place, check whether any visibility rules restrict access to specific groups.
Day-to-day
Does Live Translate store conversations?
No conversation record is kept. Each turn is transcribed, translated, and spoken back in real time, the transcript appears on screen only while the conversation is open, and there is no transcript history to open afterwards. On a shared phone in one device mode, tap Clear when you’re done so the next person who picks it up doesn’t see the conversation; in two devices mode, Leave session clears it.
Two qualifications, both worth knowing if you are answering a compliance question:
- Tapping Wrong? on a turn deliberately saves that phrase and its shown translation to a review queue at Admin → Translation, so an admin can correct the glossary. This is the only durable copy.
- In two devices mode, each turn travels to the other phone over MangoApps’ shared real-time message service, labelled with your organization’s ID, and is deleted within roughly 90 minutes (a cleanup job runs every 30 minutes). Until then it holds the turn’s text and audio. No screen in the product reads it — reaching it needs direct database access — and one device mode never uses it.
How does the “Detect language” option work?
In one-device mode, either side can be set to Detect language instead of a specific language. The app identifies the speaker’s language automatically from their speech across all 15 supported languages. The side set to Detect must speak first, and both sides can’t be set to Detect at once. In two-device (paired) mode, Detect is not available because each person’s selection becomes the other person’s translation target, so both must be set to a specific language.
How do I add glossary terms in bulk?
Go to Admin → Translation, choose the CSV file in the Terminology section, then click Import CSV. The file uses the same columns Export CSV writes, under this header row:
source_language,target_language,source_text,target_text,term_type,case_sensitive,active
term_type is mapping or do_not_translate. The last two columns are optional — leave them out or blank and a term imports as case-insensitive and active; otherwise put true or false. Save the file as CSV UTF-8 with commas between columns.
The file must be under 1 MB and contain no more than 2,000 rows. Import only adds new terms: a row with the same source language, target language, and source text as an existing entry is skipped, even if you changed its translation in the file — edit existing terms in the list instead. To see the expected format, export your current glossary first.
What are the industry starter packs?
Three pre-built terminology sets you can adopt with one click at Admin → Translation:
- Warehouse & Manufacturing Safety (English → Spanish) — 13 terms including lockout-tagout, PPE, and muster point
- Retail & Store Operations (English → Spanish) — 9 terms including stockroom, planogram, and shrinkage
- Healthcare & Facilities (English → Spanish) — 7 terms including shift handoff, charting, and isolation room
Adopting a pack clones its terms into your glossary. You can edit or remove them freely afterward. A pack shows as already adopted once all of its terms exist in your glossary.
How do I review flagged translations?
Go to Admin → Translation. The Flagged translations section shows pending flags, oldest first (up to 50 displayed; the badge shows the true total). For each flag you can type the word that came out wrong into the empty Term box (the flagged sentence is shown above it as Original:), enter what it should translate to, and click Add term — which creates a glossary mapping and closes the flag in one step — or click Dismiss to close it without a term. If the card says a term can’t fix it — the language wasn’t recognised, or both sides were the same language — dismiss it instead. Dismissing doesn’t delete the flag: its phrase still counts in the most-flagged phrases on the Analytics page. You receive an in-app notification when flags arrive (at most once per 6 hours) and an email digest (at most once per 24 hours).
Can glossary terms be case-sensitive?
Yes. Each term has a Case-sensitive option you can toggle when creating or editing it. When enabled, “PPE” matches only “PPE”, not “ppe” or “Ppe”. The default is case-insensitive matching.
When something looks wrong
“The app said it couldn’t make out any speech”
The app detected no speech in the recording. Common causes: a tap instead of a hold (the minimum hold is about 350 milliseconds), speaking too far from the microphone, or the microphone being muted at the OS level. In hands-free mode, an utterance shorter than 350 milliseconds is also discarded as a non-speech sound.
“Microphone access is blocked”
The browser refused Live Translate the microphone, usually because microphone permission is turned off for your MangoApps site. Allow the microphone in the browser’s site settings — on a phone, also check that the browser or app is allowed to use the microphone in the phone’s settings — then try again. “No microphone was found on this device.” and “The microphone is being used by another app — close it and try again.” are the other two microphone messages; each says what to fix.
“This browser can’t use the microphone here. Open the page over https, or try another browser.” is different: the page isn’t on a secure (https) connection, or the browser or in-app view can’t record audio. Open the site over https, or switch browsers.
“Couldn’t play automatically — tap Play again on the last message.”
The browser blocked the translation from playing on its own. The translation is still on screen; tap Play again on that turn to hear it. Play again stays available on the newest 10 turns.
“Translation didn’t go through — please try again.”
Speech recognition or the translation provider failed on that turn, most often a temporary provider problem. Try the turn again. If it keeps happening, check the Translation engine at the top of Admin → Translation: on MangoApps’ managed engine (no provider of your own connected), contact MangoApps support. A provider your organization connected itself shows its own message instead — “Translation isn’t set up correctly” or a usage-limit message — and the fix is in that provider’s integration.
“Taking too long — try again.”
On one device, a turn that hasn’t come back within 60 seconds is abandoned. Try again, keeping the turn short.
“You’ve reached the hourly limit for this feature”
Each person can make 200 interpreter attempts an hour, shared between one-device and two-device mode. Every attempt that reaches the server counts — including one where no speech was heard — and the count resets at the top of each hour, UTC. Starting a two-device session has its own limit of 60 an hour. Both limits are per signed-in account, not per device or per session — so on one device, both speakers’ turns count against the account that is signed in.
“The translation used the wrong word for our equipment”
The translation engine may not know your company’s vocabulary. Add the correct term at Admin → Translation: create a Mapping with the source word and its correct translation, or a Do not translate rule for terms like brand names or product codes that should pass through untouched. The glossary applies on the very next turn with no restart or delay.
“My partner can’t join with the session code”
A paired session code is 6 characters (letters and numbers) and expires after 2 hours. New codes leave out the letters O and I, so typing O for a zero or I for a one still finds the session. Only the first person to enter the code can claim the second seat — a third person is rejected even with the correct code. The joiner must also belong to the same organization, have access to Live Translate there, and join from that organization’s own site. After 20 wrong codes in 10 minutes, that person can’t join any session for up to 10 minutes. If joining still fails, have the host start a new session and share the fresh code.
“The other person left.”
The full message reads “The other person left. If they don’t come back, leave and start a new session.” The other phone left the conversation — they pressed Leave, closed or navigated away from the page, lost their connection, or lost access to Live Translate. If hands-free was on, it switches off. The session stays open for the rest of its 2 hours, and only that same person can rejoin it with the same code — as long as they still have access. Otherwise, start a new session.
“The session ended or expired. To keep talking, one of you starts a new session and the other joins with its code.”
Your turn was sent to a session that no longer accepts it: sessions last 2 hours from when they were started, and a session also refuses someone who can no longer take part in it. One of you starts a new session and reads or shows its code; the other joins with it. Both pressing Start makes two separate sessions, each waiting for the other.
“Live Translate isn’t available right now — the speech service isn’t set up on this platform.”
This notice appears only when the platform’s OpenAI connection — which runs speech-to-text and text-to-speech for every turn — isn’t set up, and the app’s buttons are disabled while it shows. Waiting won’t clear it; contact MangoApps support. An outage looks different: there is no notice, and turns say “Translation didn’t go through — please try again.”
“You no longer have access to Live Translate.” / “Live Translate isn’t available for your account.”
You see one of these when you try a turn or start a session on a page you opened earlier: the first when you aren’t in Live Translate’s audience any more — for example it was made admin-only, or your group lost access — and the second when your membership in this organization was deactivated. If Live Translate was turned off for the whole organization, the turn says “Live Translate was turned off for your organization.” Hands-free switches off on any of the three. If you’re outside the audience, reopening the page shows “You don’t have access to Live Translate. Ask your administrator to make it available to you.” An admin can publish Live Translate to you or your group, or reactivate your membership.
“I can’t see the Company terminology button”
The Company terminology button appears only for admins and super admins. It opens Admin → Translation (the page titled “Translation”), where the glossary lives in the Terminology section. Employees and managers see the interpreter but not the button — this is by design, since only admins can manage terminology.
Licensing and limits
Does Live Translate require a license?
No. Live Translate is pay-as-you-go with no license to purchase. An admin enables it at Admin → Apps → All Apps, and each turn’s usage — speech-to-text, translation, and text-to-speech — draws your organization’s AI credits (Billing & Payments → AI). The one exception is translation on a provider your organization connected itself (its own AWS Translate or DeepL account): that provider bills your account directly, so those characters draw no credits. By default turns keep running when the credits are used up, and what you use past your balance is invoiced. A turn is refused, with the limit’s own message instead of a “try again”, only when the credits are used up and Keep AI running past your credit balance is off (or your account is set to stop at zero), or when your organization’s monthly AI budget or cap, or a person’s own monthly AI limit, is reached. A new limit can take a few minutes to apply.
What are the limits I should know about?
| Limit | Value |
|---|---|
| Supported languages | 15, plus Detect in one-device mode |
| Turns per user per hour | 200 |
| Paired-session creates per user per hour | 60 |
| Maximum audio upload per turn | 2 MB |
| Maximum hands-free utterance | 15 seconds |
| Paired session expiry | 2 hours |
| Participants per paired session | 2 (host + one joiner) |
| Glossary CSV import file size | 1 MB |
| Glossary CSV import rows | 2,000 |
More help
- Live Translate Overview
- Ask AI — the assistant answers Live Translate questions from these articles.