Skip to main content
Gateway routing is not available on all plans. Contact us at [email protected] to activate the feature and plug in your LiteLLM gateway endpoint for your organization.
Instead of calling model providers (OpenAI, Anthropic, Gemini, Mistral) directly, Clarifeye can send its LLM and agentic calls through an OpenAI-compatible LiteLLM gateway that you or your operator control. This lets you centralize provider keys, spend, logging, and access policy in one place, while Clarifeye keeps working exactly as before.

​
What can be routed

Once the gateway is enabled for your instance, you opt in per model — there is no global switch. Any model that Clarifeye lets you name can go through the gateway, including:
  • The Playground assistant (chat, agentic tool use, and playbook execution).
  • Extractors: object extraction, tag extraction, and document-tag extraction.
Models you leave unprefixed keep calling their provider directly, so you can move one extractor or one Playground agent onto the gateway without touching the rest.

​
Set the model name

You choose the gateway by the value you type into the model field — the same field where you’d normally enter a model such as gpt-5 or anthropic/claude-opus-4-8. To route a model through the gateway, prefix its name with lite-llm-gateway/:
Clarifeye strips the lite-llm-gateway/ prefix and passes the remainder (<provider>/<model>) to the gateway verbatim, so it must match a model your gateway is configured to serve. For reasoning models, add a reasoning-effort suffix after a colon (:low, :medium, :high), exactly as you would without the gateway:

​
Examples

With a reasoning-effort suffix:
Whitespace around the value is ignored, so a stray leading or trailing space won’t accidentally bypass gateway routing.