AIGINT Gateway — LLM routing that picks the model

Gateway scores every request by domain, difficulty, and budget, then sends it to the model that fits. A hard question escalates into a multi-agent debate that returns a Conviction-scored verdict with its sources.

There are two ways in: the Reasoning Console, a pay-per-use chat, and the Gateway API, an OpenAI-compatible endpoint for your own apps.

• Model selection scored per request by the MCI router
• Hard questions escalate to the full multi-agent panel
• OpenAI-compatible API you can point an existing client at
• Pay for what you use, with the first answer free
• Energy cost and CO₂ shown on every request