Forge is preparing the requested surface and verifying the live route.
Forge is preparing the requested surface and verifying the live route.
Forge Routing is the decision layer that chooses how each request moves through providers, engines, and budgets.
Routing is the part of Forge that decides which execution path should handle a request. It is not just provider failover. It combines intent classification, pricing data, live health, and account policy so the platform can pick the right path automatically.
That routing layer is bundled inside paid Forge subscriptions, but it also underpins the direct tool and engine access model for external buyers who only want metered usage.
These are the high-value pieces of Forge Routing that are now surfaced as an actual product page instead of a dead link.
Forge classifies requests before model execution so it can choose the best provider path for cost, quality, latency, and safety instead of sending every request through a fixed model.
Routing policies can bias toward the cheapest viable model, the fastest viable model, or a quality-first path, while preserving failover and provider health checks.
Routing sits behind the full Forge security, memory, and validation stack, so customers get more than simple provider switching.
Routing follows the same commercial model as the rest of the managed Forge layer.
Routing is part of every paid Forge subscription, which means subscribers get the managed decision layer without buying it as a separate bundle.
Outside agents and enterprise accounts can still pay for routing-backed execution by quote, using machine-payment rails instead of a full Forge subscription.
The same routing primitives power tools, engine access, and higher-level products, so the commercial story stays consistent across the platform.
See how routing participates in Forge subscriptions and pay-per-use engine access.