Portkey AI Gateway Complete Guide: Enterprise Unified Calling for 250+ Models
TL;DR: Portkey AI Gateway is an enterprise unified gateway (250+ models) adding semantic cache, Guardrails, retries and full observability on top of OpenAI compatibility — built for enterprise governance and cost control.
Why Portkey?
- 250+ models: OpenAI/Anthropic/Google/DeepSeek/Groq/Cohere/Bedrock/Vertex, one SDK.
- Semantic cache: identical/near-identical prompts hit cache, 30-50% free quota saved.
- Guardrails: PII, content moderation, cost/latency SLO, auto block.
- Retries: exponential backoff + cross-provider circuit break.
- Observability: per-request logs, token cost, latency, success, export to Datadog/Grafana.
Quick Start
npm i -g @portkey-ai/gateway
# or Portkey Cloud: https://portkey.ai
from openai import OpenAI
client = OpenAI(base_url="https://api.portkey.ai/v1", api_key="PORTKEY_API_KEY",
default_headers={"x-portkey-provider":"openai","x-portkey-api-key":"sk-xxx"})
client.chat.completions.create(model="gpt-4o-mini", messages=[{"role":"user","content":"hello"}])
Fallback:
{"x-portkey-strategy":{"mode":"fallback","providers":["openai","deepseek","groq"]}}
Unified Calling
- Cache: high-frequency prompts hit cache, 30% cost cut at 20-40% hit rate.
- Guardrails:
pii: block+cost: max $0.01/reqauto reject. - Observability: drill by provider/model/key, alert on p95>2s or success<99%.
vs Others
| Solution | Models | Form | Best For |
|---|---|---|---|
| Portkey | 250+ | Cloud + self-host | Enterprise |
| OneAPI | 10+ | Go binary | Self-host + monetize |
| LiteLLM | 100+ | Python | Python teams |
| Cherry | 300+ | Desktop | Personal |
Best Practices
- Free-first + cache, paid as fallback.
- Hard budgets per team.
- Canary new providers at 1%.
Summary
- Enterprise: 250+ + cache + Guardrails + observability.
- Cost: cache + free-first = 30-50% cut.
- Stack with OneAPI: upstream → OneAPI
mymodel.
🚀 Get Started: One-Click Free API Access
Want to call all the free models above with a single API key, no need to sign up for each provider? Apishare.cc provides a unified API Key — one key, 100+ models, free models at zero cost.
👉 Register on Apishare.cc → Get your unified API Key
📊 Want to see more free model rankings? Check out the Sep 2026 Free LLM API Rankings →
About the Free API Aggregator
The models covered in this guide are all served through the APIShare free API aggregator, which gives you one key for the whole catalog.
- Full model catalog: APIShare free API directory
- Sign up for a free trial key: Register and claim your API key