Skip to main content
Portkey works with any tool or application that supports OpenAI-compatible APIs. Add enterprise features—observability, reliability, cost controls—with just 2 configuration changes.

​
Quick Start

Find your project’s OpenAI settings and update:
  1. Base URL: https://api.portkey.ai/v1
  2. API Key: Your Portkey API key
That’s it! Your app now routes through Portkey.

All requests appear in Portkey logs

You now get:
  • ✅ Full observability (costs, latency, logs)
  • ✅ Access to 250+ LLM providers
  • ✅ Automatic fallbacks and retries
  • ✅ Budget controls per team/project

​
Why Add Portkey?

Enterprise Observability

Every request logged with costs, latency, tokens. Track usage across teams.

Multi-Provider Access

Switch between OpenAI, Anthropic, Google, and 250+ models without code changes.

Production Reliability

Automatic fallbacks, retries, load balancing—configured once, works everywhere.

Cost & Access Control

Budget limits per team. Rate limiting. Centralized credential management.

​
Setup

​
1. Add Provider in Model Catalog

  1. Go to Model Catalog → Add Provider
  2. Select your provider (OpenAI, Anthropic, Google, etc.)
  3. Choose existing credentials or create new by entering your API keys
  4. Name your provider (e.g., openai-prod)
Your provider slug will be @openai-prod (or whatever you named it).

Complete Model Catalog Guide →

Set up budgets, rate limits, and manage credentials

​
2. Get Portkey API Key

Create your Portkey API key at app.portkey.ai/api-keys

​
3. Configure Your Application

Most OpenAI-compatible apps have settings for: Base URL / Endpoint 
https://api.portkey.ai/v1
API Key 
Your Portkey API Key (from step 2)
Model (if configurable) 
@openai-prod/gpt-4o
If your app requires the OpenAI format (just model name), use a Portkey config with your default model.

​
Common Integration Patterns

If your app allows custom base URL and API key: 
Base URL: https://api.portkey.ai/v1
API Key: PORTKEY_API_KEY
Model: @openai-prod/gpt-4o

​
Pattern 2: With Config

If your app only accepts model names like gpt-4o:
  1. Create a config in Portkey dashboard:
{
  "override_params": {
    "model": "@openai-prod/gpt-4o"
  }
}
  1. Use the config in your app:
Base URL: https://api.portkey.ai/v1
API Key: PORTKEY_API_KEY
Model: gpt-4o  (the config will override this)
Add config to API key defaults or pass via header.

​
Pattern 3: Environment Variables

Many apps use environment variables: 
OPENAI_API_BASE=https://api.portkey.ai/v1
OPENAI_API_KEY=PORTKEY_API_KEY
OPENAI_MODEL=@openai-prod/gpt-4o

​
Switching Providers

Change the model string to switch providers: 
@openai-prod/gpt-4o        # OpenAI
@anthropic-prod/claude-sonnet-4    # Anthropic
@google-prod/gemini-2.0-flash      # Google
All without changing your application code!

​
Advanced Features via Configs

For production features like fallbacks, caching, and load balancing:
  1. Create a config in Portkey dashboard
  2. Attach it to your API key defaults
  3. Or pass via header: x-portkey-config: your-config-id
Example config with fallbacks: 
{
  "strategy": {"mode": "fallback"},
  "targets": [
    {"override_params": {"model": "@openai-prod/gpt-4o"}},
    {"override_params": {"model": "@anthropic-prod/claude-sonnet-4"}}
  ]
}

Learn About Configs →

Fallbacks, retries, caching, load balancing, and more

​
Enterprise Governance

For teams and organizations:

​
Budget Controls

Set budget limits per team or project:
  1. Go to Model Catalog
  2. Create separate providers for each team
  3. Set budget and rate limits
  4. Distribute team-specific API keys

Budget Limits Guide →

Set up cost controls and alerts

​
Access Management

Control who can access which models:
  • Model provisioning - Enable/disable models per provider
  • API key scopes - Limit what each key can do
  • Team workspaces - Isolate teams with separate workspaces

Access Control Guide →

Set up RBAC and team permissions

​
Usage Tracking

Track usage by team, project, or user:
  • Add metadata to API keys
  • Filter logs by metadata tags
  • Monitor costs per team
  • Set up alerts for unusual usage

Metadata Guide →

Track and filter by custom tags

​
Common Applications

Portkey works with: ✅ AI Tools - Cursor, Windsurf, Cline, Continue
✅ Chat UIs - Open WebUI, LibreChat, AnythingLLM
✅ Automation - n8n, Make, Zapier
✅ No-Code - LangFlow, Flowise, Dify
✅ Frameworks - Any OpenAI-compatible SDK
✅ Custom Apps - Your own applications See specific integration guides:

Cursor

AI code editor integration

Cline

VS Code AI assistant

Open WebUI

Self-hosted chat interface

n8n

Workflow automation

​
Troubleshooting

​
Can’t find OpenAI settings?

Look for sections labeled:
  • “OpenAI Compatible”
  • “Custom Endpoint”
  • “API Configuration”
  • “LLM Settings”
  • “Model Provider”

​
Model not working?

Make sure you’re using the full provider slug:
  • ✅ @openai-prod/gpt-4o (correct)
  • ❌ gpt-4o (needs provider slug)
Or use a config to set the default model.

​
Getting authentication errors?

  1. Verify your Portkey API key is correct
  2. Check that base URL is exactly: https://api.portkey.ai/v1
  3. Make sure provider slug matches what’s in Model Catalog

​
Next Steps

Model Catalog

Set up providers and budgets

Configs

Configure fallbacks and routing

Observability

Track costs and performance

Guardrails

Add PII detection and filtering
Questions? Join our Discord Community or check out more integrations.