日本語 
FlashLabs has introduced Routing DSL, a new capability for OrcaRouter, the AI inference gateway developed by Continuum AI and distributed exclusively in Japan by the company. This feature lets organizations sorta design and manage multi model AI workflows using declarative YAML plus CEL based inference graphs, kind of in one place.
The launch is happening because enterprises are increasingly focusing on mixing AI models rather than trusting only one model for every task. Even though frontier models bring really strong reasoning abilities, using them for everyday chores can bump costs, and that’s not great. Routing DSL addresses this by directing workloads to different models based on factors such as task type, complexity, quality requirements, cost, and latency.
The system supports five routing approaches, including difficulty-based routing, task-based model selection, parallel execution across multiple models, automated fallbacks, and policy-based optimization. This enables organizations to reserve advanced models for complex reasoning tasks while using lower-cost models for simpler operations.
Also Read: TCS and Anthropic Join Forces to Move Enterprise AI Beyond the Pilot Stage
FlashLabs says this approach can help companies get high quality AI results without having to rely only on those premium models. Since the routing logic is written in YAML, teams can adjust workflows without rewriting the application code, and still keep clear visibility into how the requests get routed, when a fallback happens, and which models are actually picked.
This feature also plugs into OrcaRouter’s existing routing and failover abilities, and it can open access to more than 200 AI models through one single platform, so you don’t have to juggle things elsewhere.
English