Skip to main content
The Profy Developer Platform provides a complete set of APIs that let you integrate AI Expert capabilities into your own applications. Whether you want to call an AI Agent from a backend service or build an AI-powered application for end users, the Profy platform has you covered.

Capabilities at a Glance

Agents Run

Invoke a Profy Expert (AI Agent) and receive streaming responses via SSE. Supports multi-turn conversations, tool calls, memory, and more.

Chat Completions

OpenAI-compatible chat completions. Supports profy-* model names to call Experts directly — zero code changes with the OpenAI SDK.

Events

Report custom billing events for per-use billing scenarios in your App. Includes idempotency support to prevent duplicate charges.

Models

Retrieve the list of available models and their capabilities to help you choose the right one.

Meters

Query your App’s billing meter configuration and understand the charge rules for each event.

Two Authentication Methods

The Profy platform offers two authentication methods for different scenarios:
Use cases: Server-side direct calls, scripts, background tasks
  • Prefixed with sk-pro-
  • Passed in the request header via Authorization: Bearer sk-pro-xxxx
  • Credits deducted from the developer’s own account
  • Created and managed in the Platform Console
For details, see the Authentication Guide.

Core Concepts

Expert (AI Agent)

Expert is the core product type on the Profy platform — an AI Agent equipped with a persona, tools, memory, and skills. Through the /v1/agents/run endpoint, you can invoke any published Expert via API.

Expert-as-Model

Use the /v1/chat/completions endpoint with model="profy-<identifier>" to call any published Profy Expert using the standard OpenAI SDK. The Expert’s full Agent capabilities (persona, tools, memory, skills) are completely transparent to developers — you interact with it just like calling gpt-4o. This lets creators’ domain-specific AI Agents seamlessly integrate into any tool chain that supports the OpenAI API (LangChain, Cursor, etc.).

App (External Application)

An App is an external application that integrates with Profy’s identity and billing system via OAuth + Events API. App developers can leverage Profy’s user system and credit-based billing to quickly monetize their products.

Billing Models

Developer Console

Platform Console is your unified console for managing API integrations:
  • API Keys: Create, view, and revoke API Keys
  • Usage Stats: View call volume trends and distribution
  • Call Logs: Search and inspect details of every API call
  • Playground: Test APIs online (coming soon)
The Platform Console shares login sessions with the main site app.profy.cn — simply log in to the main site and you can access it directly.

SDKs

Profy provides official SDKs for Python and TypeScript, handling authentication, streaming, error handling, and more:

MCP Server

Profy provides an MCP (Model Context Protocol) Server that works directly in IDEs like Cursor and Codex. With MCP, you can create, configure, publish, and view earnings for Experts — all without leaving your editor. For details, see MCP Server Integration.

Next Steps

Quickstart

Complete your first API call in 5 minutes

Authentication Guide

Learn about API Key and OAuth usage in detail

API Reference

Browse the complete API endpoint documentation

Build an App

Learn how to build a Profy App