Skip to main content
The ForceAI Gateway is an OpenAI-compatible control plane for LLMs and agents. These guides cover the two things you do most: connecting models, and connecting agents

LLM providers

Onboard an existing provider

Add a model for a provider ForceAI already supports (OpenAI, Anthropic, Bedrock, and many more). No code, done in the dashboard or one API call

Add a new provider

Make a provider ForceAI does not ship yet a first-class option. A short code change for any OpenAI-compatible API

A2A agents

Register an external agent as an A2A endpoint in the gateway, then call it through the same governed, logged, key-scoped surface as your models

Snowflake Cortex

A Cortex Agent, over A2A, with PAT or key-pair JWT auth

AWS Bedrock AgentCore

A Bedrock AgentCore runtime, over native A2A with SigV4

Microsoft 365 Copilot

M365 Copilot, over A2A with Entra OAuth
Every model and agent you onboard is governed the same way: scoped virtual keys, budgets and rate limits, guardrails, and full request logging. See the Smart Router guide for how agents differ from models, and how routing works