OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI APIs & Platforms

Gateway

Gateway is an open-source TypeScript AI gateway offering a unified API for routing requests to multiple language, vision, audio and image models. It also integrates AI guardrails and identifies MCP client and gateway use cases.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Gateway — OpenViST project profile in AI APIs & Platforms.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Gateway when an application needs a provider-routing layer, centralized request configuration or guardrails across multiple AI integrations. The documented quickstart is also suited to testing a local gateway before choosing a deployment path.

Who it is for

  • Developers integrating several model providers into an application
  • Teams building an internal LLM operations layer
  • Builders needing configurable routing, reliability controls or request guardrails
  • Teams exploring MCP-connected AI service access

Documented uses

  • Route application requests through one API across multiple model providers
  • Add retries, fallbacks, load balancing or conditional routing
  • Apply guardrail checks to requests or outputs
  • Expose a local gateway and console for development
Before you choose

Requirements & limits

The project is publicly available as source code under the MIT license. The documented local setup requires Node.js and npm; requests require a provider API key.

  • Whether a unified API can reduce separate provider integrations in the application
  • Whether retries, fallbacks, load balancing or conditional routing match the required request behavior
  • Whether integrated guardrails belong in the gateway request path
  • Whether the team prefers a local, server or cloud deployment path
  • Whether the application uses one of the documented client or framework integrations

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Run Gateway locally and send a first model request

Set up the local gateway, inspect its console, and make a documented request through a provider.

Before you start

  • Node.js and npm
  • A provider API key
  • Python with the documented Portkey client example

Steps

  1. Run the command below.

    Documented command
    npx @portkey-ai/gateway

  2. Use the gateway endpoint at http://localhost:8787/v1 and the console at http://localhost:8787/public/.

  3. Create a Portkey client with a provider and provider API key.

  4. Call client.chat.completions.create with messages and a model name.

Expected outcome

A local Gateway instance is running and the example application sends a model request through its unified API.

Guide limitations

  • The example requires a provider API key.
  • The documented request example uses the Python client and a chat completion call.
  • This recipe covers the local quickstart, not production deployment.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
1 October 2026
Original website title
Portkey-AI/gateway · Official homepage title · retained in its source language
How we review sourcesSuggest a correction