IO Intelligence × Portkey Integration Guide

Overview

IO Intelligence is now available as a provider within the Portkey AI Gateway.
This integration allows you to access 15+ high-performance, open-source LLMs through an OpenAI-compatible API, while leveraging Portkey’s enterprise-grade gateway capabilities such as observability, fallbacks, caching, and load balancing. With this setup, you can switch between providers without changing client-side code and maintain a unified request and response schema across your AI stack.

Key Capabilities

The main capabilities of this integration include:

Chat Completions
Model Response API (createModelResponse)
OpenAI-compatible request/response format
Bearer token authentication
Unified provider routing via Portkey

Benefits

The primary benefits of this integration include:

Unified interface across multiple AI providers.
Built-in fallback and load balancing for high availability.
Built-in observability with real-time metrics and monitoring
Automatic retry handling to improve reliability under transient failures.
Request and response logging for debugging and auditability
Easy multi-provider setup without changing client-side code
Cost tracking and optimization via intelligent routing.
Enhanced security and compliance, with centralized key management.

Integration Architecture

Client
   |
   v
Portkey Gateway (provider: "iointelligence")
   |
   v
IO Intelligence API (io.net)

Portkey acts as an intelligent routing layer between the client and IO Intelligence. Clients interact with a single, consistent API, while Portkey transparently manages provider selection, request formatting, and response normalization when communicating with IO Intelligence.

Prerequisites

Before you begin, ensure that you have:

An io.net account
A valid IO Intelligence API key
A working Portkey AI Gateway setup
Node.js ≥ 18

Environment Variables

Add the following to your .env file:

IO_INTELLIGENCE_API_KEY=your_api_key_here

Your API key must allow access to the following APIs:

/chat/completions

Quick Start

Step 1: Configure IO Intelligence in Portkey

You can add IO Intelligence as a provider in Portkey from the Portkey Dashboard or via API.

Portkey Dashboard
Via API

Option A: Configuring IO Intelligence from the Portkey Dashboard

Navigate to Providers > Add Provider

Select io.net (IO Intelligence) in the Providers List

Enter your IO Intelligence API Key

Save your Configuration

The provider will be saved internally as:

"provider": "iointelligence"

Option B: Configure IO Intelligence via APIWhen using Portkey programmatically, reference IO Intelligence as an OpenAI-compatible provider:

{
  "provider": "iointelligence",
  "apiKey": "<IO_INTELLIGENCE_API_KEY>"
}

With this, portkey automatically handles:

Base URL resolution (https://api.intelligence.io.solutions/api/v1)
Bearer authentication headers
Endpoint mapping for chat completions

Step 2: Make Your First Request

Test your integration using the following examples. Chat Completion Example:

{
  "provider": "iointelligence",
  "model": "meta-llama/Llama-3.3-70B-Instruct",
  "messages": [
    { "role": "user", "content": "Hello!" }
  ]
}

Expected Response:

{
  "provider": "iointelligence",
  "id": "chatcmpl-xyz",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Hi there!"
      }
    }
  ]
}

All responses follow the OpenAI response schema, enabling drop-in compatibility with existing tooling.

Advanced Configuration

Using IO Intelligence as Primary with Fallback

As a standard Portkey provider, IO Intelligence can be included in routing and fallback configurations together with other providers. All fallback logic is defined and controlled by the Portkey gateway. Portkey can route requests to your chosen fallback provider if the following errors are encountered:

Rate limited (Error 429)
Provider errors (Error 5xx)

Supported Endpoints

Endpoint	Function Name	Endpoint Path
Chat Completions	`chatComplete`	`/chat/completions`
Create Model Response	`createModelResponse`	`/chat/completions`

All endpoints are OpenAI-compatible and normalized by Portkey’s response transformers

Error Handling

The following table shows errors you may encounter:

Status Code	Description
401	Missing or invalid API key
404	Incorrect endpoint or invalid model
422	Invalid request payload
429	Rate limit exceeded
500	Provider-side error

Troubleshooting Checklist

Confirm .env variables are loaded.
Verify model names are correct.
Validate request payload structure.
Retry 429 / 500 errors with backoff.
Inspect Portkey gateway logs for normalized requests.

Getting Started with IO Intelligence

IO Intelligence Payments

Confidential Inference

Integrations

IO Intelligence × Portkey Integration Guide

Overview

Key Capabilities

Benefits

Integration Architecture

Prerequisites

Environment Variables

Quick Start

Step 1: Configure IO Intelligence in Portkey

Step 2: Make Your First Request

Advanced Configuration

Using IO Intelligence as Primary with Fallback

Supported Endpoints

Error Handling

Troubleshooting Checklist

​Overview

​Key Capabilities

​Benefits

​Integration Architecture

​Prerequisites

​Environment Variables

​Quick Start

​Step 1: Configure IO Intelligence in Portkey

​Step 2: Make Your First Request

​Advanced Configuration

​Using IO Intelligence as Primary with Fallback

​Supported Endpoints

​Error Handling

​Troubleshooting Checklist

Overview

Key Capabilities

Benefits

Integration Architecture

Prerequisites

Environment Variables

Quick Start

Step 1: Configure IO Intelligence in Portkey

Step 2: Make Your First Request

Advanced Configuration

Using IO Intelligence as Primary with Fallback

Supported Endpoints

Error Handling

Troubleshooting Checklist