Documentation

Learn how to integrate and use Arcpoint's unified AI gateway

Welcome to Arcpoint — the unified gateway for AI that gives you access to the best models from OpenAI, Anthropic, Google, and more through a single, OpenAI-compatible API.

Why Arcpoint?

• One API, All Models — Access GPT-4o, Claude, Gemini, and 100+ models through a single endpoint
• OpenAI Compatible — Use your existing OpenAI SDK code with zero changes, just update the base URL
• Powerful Guardrails — Rate limiting, cost control, content moderation built-in
• Full Observability — Track every request with detailed traces, costs, and analytics
• Deterministic Testing — Record production traffic and replay it without calling providers

Quick Start

Get started in under 2 minutes:

1. Get your API key

Sign up and create an API key from your dashboard.

2. Make your first request

curl https://api.arcpoint.ai/v1/chat/completions \
  -H "Authorization: Bearer $ARCPOINT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

3. Use with your favorite SDK

Arcpoint is fully OpenAI-compatible. Just change the base URL:

from openai import OpenAI

client = OpenAI(
    api_key="your-arcpoint-api-key",
    base_url="https://api.arcpoint.ai/v1"
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}]
)

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.ARCPOINT_API_KEY,
  baseURL: 'https://api.arcpoint.ai/v1',
});

const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Hello!' }],
});

Core Features

Multi-Provider Access

Route requests to any supported provider with a simple model change:

# OpenAI
"model": "gpt-4o"

# Anthropic
"model": "claude-3-5-sonnet"

# Google
"model": "gemini-pro"

Built-in Guardrails

Protect your applications with enterprise-grade controls:

• Rate Limiting — Per-user, per-org, or custom key-based limits
• Cost Control — Set budgets per request, hourly, daily, or monthly
• Content Moderation — Automatic content filtering
• Request Validation — Schema enforcement and input sanitization

Observability

Every request is tracked with:

• Request and response details
• Token usage and costs
• Latency metrics
• Model selection decisions
• Guardrail outcomes

Documentation Sections

Getting Started

• Authentication — API keys and security
• API Reference — Complete endpoint documentation
• Playground — Test API calls interactively

Configuration

Configure routing, guardrails, and policies with Manifests:

• Manifests — Configuration format and structure
• Pipelines — Request processing stages
• Steps Reference — Complete step documentation
• Expressions — Conditional logic
• Namespaces — Multi-tenant configuration

Common Patterns

• Model Routing — Smart model selection
• Rate Limiting — Quota management
• Cost Control — Budget management
• Retries & Fallbacks — Error handling
• Transcripts & Replay — Request capture and testing
• Experiments — A/B testing

Self Hosting

Run Arcpoint in your own infrastructure:

• Custom Instance — Dedicated managed infrastructure
• Open Source — Build and run from source
• Self Managed — Deploy with Docker, Helm, or Homebrew
• BYOC — Bring Your Own Cloud

Need Help?

• Discord — Join our community for support
• GitHub — Report issues or request features
• Email — Contact support@arcpoint.ai