The audio-guide problem has always been hardware: a cupboard of handsets, a charging rack, a member of staff issuing and collecting them, and one language per device. A self-guided tour on NearField is the same tour without the cupboard: the visitor's own phone, the language they already read, and a Marker on the wall instead of a queue at the desk.
What the Inclusion Bar gives you
Every published Experience carries a bar fixed to the bottom of the screen with three controls: Translate, Adjust and Listen. There is no plan gate on any of it. A free account gets the same bar as an Enterprise one.
The languages are exactly twenty, and one of them is English. English is the source language and the reset, so there are nineteen translation targets. In full: English, Spanish, French, German, Portuguese, Italian, Chinese (Simplified), Chinese (Traditional), Japanese, Korean, Arabic, Hindi, Russian, Dutch, Polish, Vietnamese, Thai, Turkish, Swedish and Ukrainian. Each is offered in its own script: a visitor looking for Japanese is looking for 日本語, not for the word Japanese.
Live example · Ridgewood Museum of Natural History
Open full sizeBuild one stop
- 1
Create one Experience per object, not per room
Name it after the thing: The Ridgewood Mammoth, not Gallery 2 Panel 4. The name is what you will see in your own stats for the rest of the deployment.
- 2
Add a Heading and a Text block
Aim for 120 to 180 words of body text. That is about a minute of read-aloud, which is roughly how long someone stands in front of one object.
- 3
Add an Image, and fill in its Alt text
A detail shot of the thing they are looking at. The alt text is translated along with everything else, so it is doing real work here.
- 4
Add an Audio block if you have a recording
Upload the file and paste the script into Transcript (optional). The transcript renders behind a Transcript summary the visitor can open, and because it is a text field it translates.
- 5
Set the voice in the Settings tab
Text-to-speech voice is per Experience: Off, Adult Female, Child Female, Adult Male, Child Male. Off hides the Listen control entirely on that page.
- 6
Add an Explore More block at the bottom
It lists your other published Experiences and badges the ones this visitor has already opened, which turns each stop into a way of finding the next one.
- 7
Publish, then link a Marker
Mount it beside the existing label at a consistent height, so the second stop teaches nobody anything new about where to look.
Writing for a page that will be read aloud
Read-aloud narrates what is on the screen, top to bottom, in whatever language is currently selected. That single fact should change how you write.
- Block order is the script. If the caption should come after the story, put it after the story.
- Write sentences that survive without the picture. To the left of the tusk means nothing to somebody listening with the phone in their pocket.
- Never put words inside an image. Text baked into a graphic is not translated, not read aloud, and not resizable by the text-size control.
- Say the number. A date, a weight, a depth read aloud is worth three adjectives.
- Weekly opening hours are narrated, but the live Open now status deliberately is not, because the recording is cached, and a cached open now would be wrong half the day.
What translates and what does not
| Translated | Not translated |
|---|---|
| Headings, body text, button labels, captions, accordion items, quiz and choice options, image alt text, and audio transcripts. | Words baked into an image or a graphic. |
| NearField's own interface strings: the verify prompts, the hours board, the check-in and passport labels, the consent banner. | A recording you uploaded, and a video's own soundtrack or its captions. |
| Server copy you wrote, such as a donation thank-you. | Anything inside an embedded frame: video, maps, Google Forms, social posts. |
| A PDF opened from a File Action, and the Stripe return page after a donation. |
Recorded audio and read-aloud together
Use both. A curator's own voice beats any synthetic one in the language it was recorded in, and read-aloud covers the other nineteen without you booking a studio nineteen times. The recorded tour is the Audio block; the translated tour is the same page's text.
One timing detail worth knowing: an English page with nothing personalised on it plays from a recording generated ahead of time and starts immediately. Any other language, or a page that has a visitor's own answer written into it, is synthesised on the spot and takes a moment before the playbar appears. That pause is not a fault, and it is the reason to keep a stop to a minute rather than five.
Seeing which languages people asked for
- Each Experience's report carries an Inclusion Bar panel: translations used, accessibility activated, text-to-speech plays, and Top languages requested.
- On Growth, the insight deck says it in a sentence: Visitors asked to read your page in Portuguese.
- That list is a signpost for print, not just for the app. Three seasons of Ukrainian requests is an argument for a Ukrainian line on the physical label.
Example
Ridgewood Museum of Natural History
Ridgewood runs nine stops, one Marker each, mounted at the lower left corner of every existing interpretive panel so visitors learn the position once.
Each stop is a Heading, a 160-word Text block, one detail Image with real alt text, an Audio block carrying the curator's ninety-second recording and its full transcript, and an Explore More block. The voice is set to Adult Female on every page for consistency.
In its first season the bar recorded 1,180 read-aloud plays and 640 translations, of which Spanish was over half and Ukrainian was fourth. Ridgewood added a printed Spanish line to the three busiest panels on the strength of it, and left the other six to the phone.
Can I supply my own translation of a page?
Not as an override. Translation runs live and there is no glossary or per-phrase pin. If a wording has to be exact in a particular language, build a second Experience written in that language and link to it from the first.
Do I need a paid plan for translation or read-aloud?
No. The Inclusion Bar, all twenty languages, read-aloud, text sizing and high contrast are on every plan including Free. What a paid plan adds is more live pages, more Marker placements, and the blocks that ask visitors something.
Can I turn read-aloud off on one page?
Yes. Set Text-to-speech voice to Off in that Experience's Settings tab and the Listen control disappears from that page only. Useful for a stop in a quiet gallery where a phone speaker would carry.
How long does one stop take to build?
About twenty minutes once you have the words, most of which is writing the 150 words, not using the builder. Nine stops is a working day, not a project.
The bar at the bottom of every Experience is where translation and read-aloud live. Here is every control on it.
Languages and read-aloud