runlet.
Open panel
THE AI LAYER FOR YOUR PRODUCT

Users ask. Your product acts.

Customers ask in everyday language. Runlet finds the right action using your product knowledge, and your backend carries it out.

Get started See how it works

Cloud or on your server · Panel sign-in currently uses Iranian mobile numbers

Outside Iran? Contact us about your integration
EXAMPLE EXPERIENCE IN YOUR APP A customer leaves a confusing dashboard behind, makes a simple request, and sees a completed form
CUSTOMER ASKS

Create a form titled “Partnership request”.

WITH RUNLET
AFTER YOUR BACKEND APPROVES AND EXECUTES

Partnership request form is ready.

ONE INTEGRATION. REAL ACTIONS.
OpenAPIRelevant knowledgeServer-side execution
A QUICK PREVIEW

What would your customers ask?

Choose a request. The result belongs in your app's own interface.

Illustrative preview · no live API calls
CUSTOMER ASKS

Make a form titled “Customer feedback”.

Runlet uses your OpenAPI and product knowledge to propose an action. Your backend checks permissions and executes it.

FORM CREATED

Customer feedback

NameYour nameYour feedbackHow can we improve?✓ Ready in your app

This experience, inside your product, with your own API.Connect my app
Fewer menus to navigateReuse your backend and permissionsResults in your own UI
HOW RUNLET WORKS

A conversation becomes a completed task.

Customers do not need to learn a complex panel. Runlet matches their words to your product's capabilities.

A customer's conversation is matched with product knowledge, checked as an action, and completed in the app
Customer request → relevant knowledge → approved action → result in your app
01

Show Runlet your product

Upload your OpenAPI document and product description so Runlet knows which actions exist.

02

Customers ask naturally

Runlet matches a request to relevant product knowledge and proposes one action.

03

Your app does the work

Your backend checks permissions and confirmation, executes the action, and shows the result in your own UI.

CHOOSE WHERE KNOWLEDGE LIVES

Two ways to deploy. One experience.

In both modes, AI model traffic goes through the Runlet broker.

☁MANAGED

Runlet Cloud

The panel, API catalog, descriptions, and knowledge vectors live in Runlet's cloud service.

  • Start without running a separate server
  • Manage workspaces and usage in the panel
Open panel ↗
▣YOUR SERVER

Standalone

The panel and knowledge database run on your server. Selected text for AI processing passes through the Runlet broker.

  • Deploy with Docker Compose
  • Keep the full catalog and vectors in your database
View setup ↗
CONTROL STAYS WITH YOU

AI proposes. Your system decides.

Runlet does not call your application API directly. Your backend authenticates the user, approves the action, and executes it through local handlers.

Explore the API
USAGE-BASED BILLING

Pay for model usage.

Input, output, and embedding tokens are recorded separately. View usage and calculated cost in the panel. Monthly workspace token limits help control consumption.

Choose 200,000, 500,000 or 1,000,000 tokens. Prices follow the current dollar rate.

View token packages
COMMON QUESTIONS

A few useful answers.

Does Runlet change my system directly?

No. Runlet returns a proposed action. Your backend owns access checks, confirmation, and execution.

What leaves my server in standalone mode?

Selected documentation text and user requests go through the Runlet broker for AI processing. The full catalog and vectors stay in your local database.

What do I need to integrate?

An OpenAPI document, a product description, and handlers in your backend. The panel generates a copyable integration guide.