For teams building AI into mobile apps

Ship AI features. Not your API keys.

One SDK and gateway for mobile apps. Keys stay on the server. Every request is verified, limited and costed.

Free to use during early access. Pricing will be announced before paid plans launch.

Kotlin MultiplatformSwiftFlutterReact Native
one request→ Anthropic
Device verified Key stays on server Quota OK Routed → Anthropic Cost logged $0.0004
The problem

AI in mobile apps leaks keys and money.

Your key ships inside every install.

Free users can burn your margin.

daily AI spendexample
64%

of LLM-enabled iOS apps studied leaked credentials or backend access in their traffic. 282 of 444 apps. Wake Forest University, arXiv:2606.12212, June 2026, preprint.

Features

Everything between your app and the model.

Keys never ship

Provider keys live on the gateway. Rotate without a release.

Only your real app

Play Integrity and App Attest.

Limits by plan

From RevenueCat, StoreKit or Play.

Any provider

Switch models without an app update.

Cost per user

AI cost next to what each user pays.

One SDK, every stack

Native first. Cross-platform on the same core.

AndroidiOSKotlin MultiplatformFlutterReact Native
Demo

Send a request. Watch the gateway decide.

Plan
Device
Provider
Simulation · demo prices
    0allowed
    0blocked
    $0.000spent
    Cost per user

    Know who's profitable.

    Free user

    42 requests / mo

    Revenue$0.00
    AI cost$0.21
    −$0.21/ mo

    Pro subscriber

    $4.99 · 180 requests / mo

    Revenue$4.99
    AI cost$0.62
    +$4.37/ mo

    Power user

    $4.99 · 2,400 requests / mo

    Revenue$4.99
    AI cost$7.80
    −$2.81/ mo
    Access & pricing

    Start building. We'll work out pricing with you.

    We're keeping early access free while we learn how teams use Relaykit.

    During early access
    Free for now

    Explore the gateway, test integrations and help shape the product. Paid plans may be introduced later; free early access is not a promise of permanent free use.

    Join early access
    What comes next

    Pricing is being worked out.

    We haven't finalized rates or usage allowances. We're evaluating transparent usage-based options, potentially with daily and monthly billing views.

    No surprise pricing. We'll publish the actual rates and terms before asking anyone to move to a paid plan.

    Model-provider usage is separate from Relaykit's eventual service pricing.

    FAQ

    Questions

    Can I use it today?

    We’re onboarding early-access teams in batches. Join the list to request an invitation.

    How will pricing work?

    Early access is free for now. Paid rates and allowances are not finalized; we’ll share the terms before introducing paid plans. Provider usage charges are separate.

    Do you store prompts?

    No. We keep usage metadata like user, tokens and cost. Prompts pass straight through.

    Why not Firebase AI Logic?

    It's Gemini-only. Relaykit adds OpenAI and Anthropic, plus limits tied to what users pay.

    Which platforms?

    Kotlin Multiplatform and Swift first, then Flutter and React Native.

    Stop shipping keys.

    Only early-access emails. Unsubscribe anytime.

    You're on the list

    Four taps help us build the right thing first. Optional.

    Platforms
    Providers
    Biggest problem
    Future billing preference (optional)