Built a realtime phonics + reading practice app that adapts to the reader's mistakes and reinforces sound-it-out.A little laggy because I'm running locally but so excited about this for my early reader!

Awesome. This might be helpful with my sons speech therapy

I’ll let you know when it’s ready!

@vinhocent Yessss this is amazing. Product please. Also — for apps like this where the speaker is likely to enunciate slowly I would increase the silence_duration_ms setting on VAD or try Semantic VAD, these will both increase the time before the model starts speaking.

Adaptive phonics via realtime voice is exactly the applied AI that matters. Local lag is a deployment problem, not a design limitation.

This is super cool that you can do all this with one tool now! I helped build the early MVP for back in 2023 which does this for teachers so kids get 1:1 feedback. We used Eleven Labs text-to-speech, Deepgram speech-to-text, and Claude for text processing.

Following this. I have interactions with limited literacy adults who could really use a high quality tool to help them get up to speed. Many spent time incarcerated and didn't get needed literacy intervention while inside. Early literacy books work, but they feel a bit 'child focused' for adults.

Voice for toddlers is *really* hard to get right. Kudos for trying! If it doesn't stick, you might find something here:

you'll be forever remembered for your parent helping ships

pls share github link

Omg I need this for my son - could you please share if possible? What voice model are you using?

I would love this for my sister.

Impressed you spun this up on literally day one of this release! Honest question though - why is this better than a physical phonics book?

This is reaaaally cool, is there a website to play with or a GitHub to star it?

Omg my wife is learning mandarin and the tones are nightmarish. May fork the idea for that use case

That sounds amazing! Can't wait to see it launched!

@pbbakkum bravo

That sounds like a fantastic tool for early readers! Can't wait to see how it evolves 📚✨

That's incredible! Can't wait to see how it helps kids thrive! 🎉📚

What do you use for the voice in the app? I experimented with using AI for phonics a year ago and couldn’t get the voice models to pronounce individual phonemes. Here’s what that looked like:

Waiting for the product!! Thank you!

this is what good PM + good dev alignment looks like. product sense meets real execution.

Very cool, share once done please

Brilliant use of voice AI. Teaching kids to sound-it-out with adaptive tech creates real value, not flashy features. How are you handling the local latency?

'fraaahg' lol, that would actually pass in gpt-realtime-1.5-michigan-accent. cool idea!

yo that sounds super cool ngl, can’t wait to see it in action!

🐂｡◕‿◕｡

you inspired me to make a similar app for my kid. brb

Really cool! The /r/ has a little schwar but love the idea!

would be cool for other languages

This looks cool! I’d love to give it a try (especially to work on my accent) if you end up releasing it. Thanks for sharing, Claire.

I need this for my kiddos

Nice friggin job!!!

Great idea, I'd like to do this for learning Spanish pronunciation.

adapts to mistakes smart. running local how's the inference holding up