“Trim the opening and lower the music.”
Your app’s trim and volume tools prepare an edit for review before export.
Let AI work
inside your app.
Connect AI to your app’s capabilities. AGKit Cloud powers the models, configuration, and spending controls behind your app.
AGKit is a Canadian company, with servers in Canada and the European Union.
Illustration: a video editor, a transcription and notes app, a document reader, and a voice assistant with a speaking sphere flow into one another. These are illustrative application concepts.
The AI providers to power your apps.
See capabilities →- OpenAI
- Google Gemini
- xAI
Apple
Change the model.
Keep the app.
Your choice of models, managed through AGKit Cloud. Configure providers, model routes, and media capabilities from the dashboard. Your app executes its own actions when Cloud calls its tools. Change the model without a new app release.
See what you can change →Illustration: your app sends a request to AGKit Cloud. Cloud connects to AI providers or external tools and APIs, then sends an action back to the app. The app executes its own tools and applies approved changes.
Your app
let documentEditorProfile = CloudProfile(
agentId: "document-editor",
tier: "balanced",
maxOutputTokens: 1024,
signer: BackendRouterCloudAuth(
publishableKey: "ak_pk_live_example",
userToken: "your-user-jwt"
)
)
CloudExample bindings: Claude Opus 5.5, GPT-6 Sol, or Grok 4.7, configured in Cloud for the same balanced tier.
Usage limits for
every user.
Connect your existing sign-in. AGKit Cloud verifies each user’s token, records their AI usage, and enforces per-user limits on turns, tokens, and spending.
Building a subscription app? Connect RevenueCat to require an active entitlement before a Cloud request runs.
Explore identity and usage controls →User · ****Verified identity · hashed in usage records
- Turns today
- 18 / 50
- Tokens this month
- 42,000 / 100,000
- AI spend this month
- $1.20 / $5.00
Give AI tools
inside your app.
Expose your app’s existing functions as tools the AI can call. AGKit routes calls to your app’s code and returns the results to the model, so one request can become a sequence of actions.
“Turn this recording into notes and tasks.”
Review the meeting notes and save the suggested tasks to your workspace.
“Apply these changes to the document.”
Your document tools find the right section and prepare a revision for review.
You define the tools and decide what they can do. See how to connect a tool →
Illustration: an application action, image and voice media, and an app code panel connect to a central AGKit Cloud. A signal travels out from Cloud and back along each route in turn.
App actions and media, together.
Connect your app’s tools to AI, then add image, voice, and video capabilities through Cloud’s configured routes.
Application actions
Connect creative, document, or productivity tools to a runtime that calls your application code.
Tools and permissions →Media capabilities
Generate an image, give a story a consistent narrator, or create a video through Cloud’s media routes.
Explore media →
Illustration: a proposed edit waits for review, then is approved. A separate proposal is cancelled and its draft discarded. The sequence repeats.
Keep users
in control.
Build review into actions that need it. AGKit provides supported guard and confirmation hooks; your app validates current state, applies changes, and implements undo for reversible operations.
Explore permissions and recovery →AGKit MCP
Let your coding assistant
build with AGKit.
Connect AGKit MCP to the assistant you use to code. While it helps build your app, it can create a Cloud project, configure agents and tools, set model and media routes, and manage usage limits. Review the proposed changes before they go live.
Raise the per-user monthly token limit in this project to 10 million. Proposed change: Monthly tokens per user, from 5 million to 10 million. Applies to all users in this project. After review and a separate apply: Done — the per-user monthly token limit is now 10 million. Change recorded in the audit log. Illustrative example only.
Permissions you choose
Limit access to the projects and operations your assistant needs.
Changes you can review
See the proposed changes before a plan is applied.
An audit trail
Applied plans record who made the change and what changed.
Your app.
Any screen.
Build for mobile, the browser, or your backend. Connect your app’s capabilities, then manage models, configuration, and spending through AGKit’s shared Cloud backbone.
Swift
Apple apps
Give AI access to your app’s tools and state, with an integration that fits your Swift code.
Android
Native Android
Connect AI to native app capabilities while keeping the experience in your app’s hands.
React Native
Cross-platform mobile
Add AI-powered actions to your iOS and Android apps with React Native.
Web
Web apps
Connect AI to your web app’s tools and state, with models and configuration managed in AGKit Cloud.
Imagine the possibilities.
Start with one useful action.
Choose your platform, connect a tool your app already has, and give your users a new way to use it.
Private SDK access · setup starts with your platform and application.
Get started for free.
Start with a free Cloud plan. Model usage is billed directly by your provider, with no AGKit markup.
Free$010K requests / mo
Hobby$24.99/ mo500K requests / mo
Pro$99/ mo5M requests / mo
EnterpriseCustomCustom requests / mo