Skip to content

Use Warp with the Gateway

Warp is a modern terminal with built-in AI features. Warp's custom inference endpoint setting lets you send agent requests to the Gateway instead of using Warp credits.

Prerequisites

Connect Warp to the Gateway

  1. Open Settings in Warp and search for inference endpoint
  2. Add a custom endpoint with:
    • The endpoint URL https://llm.[your-region].amazee.ai (Warp calls /v1/chat/completions on this URL)
    • Your LLM API key
    • The Gateway model IDs you want to use, e.g. claude-4-5-sonnet. See Available Models.
  3. Save, then pick one of those models in Warp's model picker

Replace [your-region] with your region code (e.g. ch103, us103).

Not the same as Bring Your Own Key

Warp's API keys (BYOK) setting only takes keys for OpenAI, Anthropic, or Google directly. Use the custom inference endpoint for the Gateway.

Requests pass through Warp

Warp sends your requests through its own servers before they reach the Gateway. That means prompts can leave your region on the way. If you picked the Gateway for data residency, check Warp's data policy first.

How it works

  • Requests to your chosen models are billed to your Gateway workspace, not to Warp credits. On Warp Business and Enterprise plans, Warp still charges its own platform credits.
  • Auto model selection always uses Warp's models and credits, and doesn't go through the Gateway. Pick a Gateway model by name.
  • Your API key is stored on your device.
  • Available on Warp's Free plan and on paid plans for individuals and companies with 10 or fewer employees. Larger companies need Warp Business or Enterprise.

Cloud Agents

Custom endpoints don't apply to Warp's Cloud Agents. The configuration is stored on your device.

Verify

Use Warp's AI features with a Gateway model selected, and check the My usage card on the AI keys page to confirm requests appear.

Troubleshooting

Auth errors or "not found"
Verify your API key is correct and the endpoint uses your key's region. If requests fail with the URL above, try https://llm.[your-region].amazee.ai/v1.
Model not available
Check that the model is available in your region. See Available Models.
Still using Warp credits
Make sure you picked a Gateway model by name, not Auto.