We're funding developers to build real apps with on-device AI. $500–$1,000 per project. Build something you'd actually ship, share what you learned, and get featured.
Two Tiers
Starter
$500 per project
Single model, one platform, focused scope.
Showcase
$1,000 per project
Full pipeline, cross-platform, complete story.
What We Ask
Not a to-do app with AI sprinkled on top. We want apps that meaningfully use on-device inference — apps you'd actually use or ship. Prototypes are fine, but they should demonstrate genuine value.
Write about what you built and how. Blog post, LinkedIn thread, video walkthrough — pick the format that fits your audience. Be honest. If something was frustrating or broken, say so.
We'll agree on milestones upfront. Week 2: working prototype. Week 4: feature-complete. Week 6: content published. 50% paid at prototype, 50% on completion.
Your experience directly shapes the SDK. File issues, tell us what's confusing, suggest improvements. This is as much a beta program as a grant.
How It Works
Apply
Short pitch — what's the app, which SDK, why it's interesting. No formal proposals.
Week 0
30 minutes to align on scope, milestones, and which tier fits. We'll set up a direct support channel.
Week 2
Core functionality running. First 50% of the grant paid at this milestone.
Week 4–6
App complete, content published, code open-sourced (if applicable). Final 50% paid. We feature your project.
What You Get
$500 or $1,000 depending on scope. Paid in two milestones.
Private channel with the Xybrid team. We'll help debug, optimize, and ship.
Your app showcased on xybrid.ai, our GitHub, and socials.
New models, features, and SDKs before they're public.
Want to present your project at a conference? We'll help with travel and submissions.
Builder badge, contributor credits, and introductions to the Xybrid network.
Project Ideas
These are starting points — bring your own ideas. The best projects are ones you're genuinely excited about.
Record, transcribe, and summarize — entirely on-device. Searchable voice journal that works on a plane.
Speak in one language, hear back in another. Whisper + LLM + Kokoro TTS across Android, iOS, and desktop.
Game characters that listen, think, and speak — no cloud APIs, no per-request costs, works offline.
Accessibility tools, creative apps, developer utilities — if it uses on-device AI in a compelling way, we want to hear about it.
Not a Fit
FAQ
No. Students, hobbyists, content creators, indie devs — all welcome. What matters is that you can ship something real within 6 weeks.
Yes. If you're integrating Xybrid into an app you're already building, that's great. We just ask that the content you create is about the AI integration, not a promo for your product.
That's expected at this stage. You'll have direct access to our team and we'll prioritize fixing issues you hit. Part of the value for us is learning where the rough edges are.
Yes. Your app is yours. We just ask for permission to feature it on our site and in our marketing. If you open-source it, we'd love to link it from our examples — but that's your choice.
Wire transfer or PayPal. 50% at working prototype (week 2), 50% on completion (week 4–6). We'll sort out the specifics on our kick-off call.
If this works, yes. First cohort is 5 spots. We'd rather support 5 builders well than spread thin across 20.
Apply
Cohort 01 has 5 spots. Tell us what you want to make.