Routific MCP. Resolve complex multi-stop routes and manage dispatch from chat.

You're connecting your AI client straight up to Routific. This isn't some gimmick; it turns your chat interface into a full-blown dispatch manager. You can build multi-stop routes, audit every outstanding order in the system, and push finalized manifests right into the driver’s mobile app—all without ever leaving the chat window.

When you need to figure out where things are going, Routific handles the complex routing math for you. If you've got a small cluster of stops that just need optimizing, you can run it immediately with solve_standalone_vrp, which calculates an optimized route right away. But what if you’re dealing with dozens of stops across multiple depots? You launch that massive job using solve_async_vrp_long.

This starts the Vehicle Routing Problem calculation in the background, giving you a unique ID; you gotta use that ID to check its status later by calling poll_async_solution until it's done. Before you even solve anything, you can generate an initial plan for the whole system using create_platform_route.

To manage the actual deliveries—the jobs themselves—you've got total control. You start a new delivery order in the platform by calling create_saas_job, where you feed it structured JSON with all the necessary order details. If an existing job changes, don't sweat it; you can modify or change details for any active record using update_saas_job.

Need to see what's pending? You run list_platform_jobs and get a comprehensive list of every outstanding delivery job on file. And if a job gets canceled before it even starts, you hit the kill switch with cancel_saas_job, which deletes and cancels that delivery job from the platform—just be warned, that action is irreversible.

Once the route is solved and the jobs are in place, you need to know the timeline. You retrieve a detailed, stop-by-stop schedule including estimated arrival times for any completed or planned run using get_route_timeline. When everything's locked down and the plan is set, you send that final manifest directly into the targeted driver’s mobile application with one command: publish_route_to_driver.

It's a simple flow. Your agent handles the complexity. You tell it what needs to happen—whether it's creating the job via structured data (create_saas_job), solving the logistics puzzle (solve_standalone_vrp), checking on its progress (poll_async_solution), or pushing the final directions out (publish_route_to_driver). You don't write complicated API calls; you just chat, and your agent does the heavy lifting.