Turn your AI agent into a self-service app

The adoption platform for AI agents. Serve 1000s of users via a simple inference API.

From the makers of an AI agent used by

Red HatTatumKeboolaApify
Automate onboarding & customization

Let users customize, group, and chain requests based on their personal or company context.

Handle diverse requests at scale

Process up to 3000 requests per second with up to 40MBs of data per each request.

Manage everything from one place

Manage users, organizations, rate-limits, scaling, timeouts, and charging from one admin panel.

All you need is an inference API - here's how it works

1

Connect inference API

Start by providing an example of your agent's interaction. Upload a sample request, the corresponding response JSON, and your OpenAPI specification. Our platform uses this blueprint to understand your agent's capabilities and automatically map its inputs and outputs.

Add Agent's API
Example Request
response.json
OpenAPI spec
Mapping
2

Set customization rules

Define the guardrails for agent customization on both a per-user and per-company level. Using a simple YAML configuration, you can control exactly what the agent can access and how users can tailor its behavior. We pre-generate this file based on your API spec to get you started instantly.

Customization Rulesyaml
# Configuration rules for agent customization
permissions:
  # Define what objects the agent can interact with
  objects:
    - companies
    - persons
    - emails
  # Specify allowed use cases
  use_cases:
    - sourcing
    - enrichment
    - predictions
  # Control access to specific data points
  datapoints:
    - contact_info.email
    - company_info.revenue
    - person_info.job_title

# Define user's ability to create new configurations
user_config:
  # 'fixed': users can only select from pre-defined templates
  # 'approval': user-created configs require admin approval
  # 'open': users can create and use new configs freely
  policy: approval
  
# Define logic for generating requests based on context
request_logic:
  - if:
      variable: user.plan
      equals: "enterprise"
    then:
      request_param:
        priority: "high"
  - if:
      variable: company.country
      in: ["USA", "CAN"]
    then:
      request_param:
        model: "0.1"
3

Customize User Experience

Control the entire look and feel of the onboarding process to create a seamless, on-brand experience. Define custom wording, apply your brand's fonts and colors, and set concurrency limits to manage resource usage.

UI Customizationjson
{
  "Wording": {
    "Activity": "Use-case",
    "Workflow": "Scenario",
    "Source": "Agent",
    "Datapoint": "Characteristic"
  },
  "Fonts": {
    "primary": "Inter, sans-serif",
    "secondary": "Poppins, sans-serif",
    "monospace": "Roboto Mono, monospace"
  },
  "Colors": {
    "primary_accent": "#2670C1",
    "background_main": "#FFFFFF"
  },
  "Concurrency": {
    "max_parallel_runs": 50
  }
}