
by Portkey
Portkey is an AI gateway giving one interface to over 250 models, with monitoring, guardrails and governance. Its docs now name it the Prisma AIRS AI Gateway.
APIFreemium
Step 1: Get a Portkey API key.
Step 2: Choose a provider and install the Python or JavaScript SDK.
Step 3: Send a chat completion through Portkey, or point your OpenAI client at the gateway.
Step 4: Review logged requests in Portkey.
Step 5: Add caching, routing or guardrails as needed.
The docs estimate 20 to 40 ms of added latency compared to direct calls, and say caching and routing can offset it.
Portkey lists ISO 27001 and SOC 2 certification, GDPR and HIPAA compliance, and AES-256 encryption in transit and at rest. Storing request bodies can be switched off on request.
No explicit timeout is imposed on free or paid plans, though the docs recommend client-side timeouts.
For enterprises, Portkey offers managed hosting inside private clouds.
Meta Model APIMeta Model API is a self-serve inference API for Meta's Muse models, covering text and agentic work, image generation, speech transcription and image segmentation. It works with existing OpenAI SDK code.