Igbo belongs in the technology we use every day.
But when we speak Igbo, many of the apps and tools we use still don't understand us. Automatic captions fail. Transcripts get the words wrong. Voice tools make us switch to English.
TokataHQ is working to change that.
Right now: getting technology to accurately hear and understand spoken Igbo.
When we speak Igbo, can the technology hear us?
Automatic Speech Recognition, or ASR, is the technology that turns spoken language into something a digital product can work with. Many everyday features depend on it.
Speaking Igbo is not the same as understanding it. A product may read Igbo text aloud and still fail to recognize someone speaking Igbo.
- Captions
- Transcription
- Voice typing
- Voice search
- Voice assistants
What this looks like in real life
Your Igbo video has no automatic captions.
You upload it to a platform or editing app. Instead of captions appearing on their own, you type every line yourself.
You speak Igbo. The transcript comes back wrong.
The words are misheard, skipped or turned into something you never said.
You have to switch to English.
Voice typing, voice search and assistants aren't much use if they only work once you stop speaking Igbo.
Why it matters
Igbo online is often spoken before it is typed: in music, videos, comedy, voice notes and family calls. When technology can't recognize speech, that is where it falls short.
- Creators
- It is more than an inconvenience. Creators may have to caption or transcribe everything themselves. Content that a platform can't understand can also be harder for audiences to find.
- Musicians
- Some features built around lyrics, transcription and search need to recognize what was sung. When they can't recognize Igbo, artists may have to supply lyrics and text themselves.
- Everyday users
- People who reach for voice typing or a voice assistant give up on Igbo, or stop using the feature.
- Families
- Children are growing up talking to devices. If those devices only understand other languages, Igbo gets left out of everyday use.
Why TokataHQ exists
A thousand people quietly fixing the same problem can make it look like no problem exists.
One creator types captions by hand. Another switches to English so a voice tool will understand. Each works around the problem alone, so the companies building these tools may never see how many people it affects.
Community experiences help us spot the patterns people otherwise solve quietly on their own.
Does it work in Igbo?
Latest: in our September 2026 tests, CapCut, Instagram and TikTok did not offer automatic captions for spoken Igbo.
What we want companies to change
- Recognize spoken Igbo accurately.
- Make automatic captions and transcription work in Igbo.
- Test with real Igbo speakers.
- Say clearly which languages each feature handles.
- Put Igbo on product roadmaps now.
In their words
“When I was doing only Igbo skits, I used to translate my skits myself.”
She opened the language menu. Other Nigerian languages were there. Igbo wasn't.
Seeing something that doesn't work in Igbo?
A caption that never appeared, a transcript full of wrong words, a voice tool that didn't understand you. Tell us. What people share helps us decide what deserves closer testing, research and advocacy.
Want to do more? Volunteer with TokataHQ. Building speech technology? Start with the Igbo data and models that already exist.