An AI that speaks — and describes the world.
Déblo is designed to work without looking at the screen. You speak naturally, Déblo answers aloud. It reads the letter you hold in your hand, describes a photo, helps with an administrative form. Compatible with your usual screen reader.
Languages and inclusion · Market 4 of 4
Accessibility audience under specification — target release 2026 / 2027 under institutional funding.
The context
Digital access is still a daily battle
Over 250 million people worldwide are blind or low-vision. Most digital interfaces remain visual-first. Screen readers exist — JAWS, NVDA, VoiceOver, TalkBack — but most consumer AI products remain barely usable without manipulating a screen.
What Déblo brings
Voice AI designed to work without a screen
Native voice control, nothing to press
Voice activation, voice conversation, voice action. No need to target a button or navigate a menu. Déblo listens continuously on demand and answers aloud.
Spoken description of images and documents
You hold a paper: invoice, prescription, administrative letter, product label. You take a photo (by long press or voice command), Déblo reads aloud, summarizes, translates or explains.
Compatible with your screen reader
JAWS on Windows, NVDA on Windows, VoiceOver on iOS and macOS, TalkBack on Android. Déblo is tested to coexist without audio collision (polite turn-taking).
No visual traps
No image-only captchas, no carousel without description, no image-button without label. Déblo public surfaces follow WCAG 2.1 AA — the minimum baseline, with WCAG 3.0 in preparation as the standard stabilizes.
Concrete use cases
Daily life, administration, work
Administrative processes
Read a letter from administration, understand a form, draft a reply, identify a reference number. Déblo guides step by step.
Daily life
Identify the content of a food label, decipher a medication leaflet, read a restaurant menu photographed on the spot, scan a bus schedule.
Training and work
Prepare a course, revise a program, write a professional email, understand a technical document. For students and professionals alike.
Technical approach
Specialized vision grounding for spoken description
Déblo integrates a vision grounding module that verbally describes the content of an image with spatial precision: where the object is, its relative size, its relation to other elements. This is what enables saying "the text is in the top right, just below the logo" rather than simply "there is some text".
Partnerships
Who we want to build this version of Déblo with
This service is not a classic commercial product. It is built in partnership with associations of blind and low-vision people, public institutions, and international funders who finance digital inclusion. We are looking for co-creators, not customers.
Associations and federations of blind and low-vision people (Federation of the Blind, Hadley Institute, Lighthouse International, African federations)
Funders and inclusion programs (UNESCO, AFD, GIZ, World Bank, Mastercard Foundation, Gates Foundation, AfDB)
Training institutions for visually impaired people
Ministries of health, social cohesion and social affairs — public partnerships
FAQ
Your questions about Déblo Accessibility
Does Déblo replace my screen reader?
What devices is it available on?
Do I have to say "OK Déblo" to activate it?
How much does the service cost?
Which languages are supported?
How do you ensure accessibility quality?
When will this be available?
Which organizations do you work with?
Let's build this version of Déblo together
If you represent an association of blind people, a public institution or a funder committed to digital inclusion, let's talk. This version of Déblo is conceived as a common good, not a closed proprietary product.
