Tell Smith how the work gets done. It designs the agents, tools, and routing to do it, then deploys the whole system, live and traced.
No SDK to learn · no infrastructure to manage · deploy in one click
Don't start with agents. Start with the work. One conversation later it's designed, deployed, and traced.
Describe the work you want done, in plain language.
Smith generates the agent graph, with routing welded in.
Tune any agent's model, tools, and prompt.
Compile and publish a versioned live endpoint.
Watch real traffic flow through the graph, traced.
No graph code, no SDK. Describe the process you run today and Smith drafts the agents, tools, and routing to run it, showing you the plan as it forges, step by step.
Open any agent to change its model, temperature, tools, and system prompt. Smith suggests; you decide. Nothing hides behind one giant prompt.
One action compiles the graph and publishes a versioned endpoint. Watch real traffic flow through it, with runs, latency, and traces built in.
Most tools stop at a single agent. Smith generates the orchestration: supervisors, specialists, parallel work, and feedback loops.
A router delegates each task to the right specialist and decides when the job is done.
Each step hands its output to the next, in a fixed order, start to finish.
Run independent agents at once, then aggregate their results into one answer.
An agent acts, a critic checks, and the work loops until it meets the bar.
Visual builders hand you a blank canvas; frameworks like LangGraph and CrewAI have you assemble the agents yourself. Smith starts from the work: it generates the graph from your description, then hands you the controls.
The whole system is a graph you can see, never a hidden chain.
Every agent, tool, and edge is inspectable and tunable, before and after deploy.
One action takes a draft to a live, versioned endpoint on the Smith runtime.
Live traces and metrics are built in, right where you built the system.
Those tools are agent orchestration first: a framework or canvas where you assemble agents, crews, and graphs yourself. Smith is workflow authoring first: you describe the work, and it generates the multi-agent graph, then lets you tune any node and deploy with tracing, in one place. The graph is an output, not a chore.
People who own a workflow: support leads, ops managers, founders, and developers who would rather skip the scaffolding. If you can describe how tickets get triaged or refunds get approved, you can commission a system that does it.
No. You describe the work in plain language and Smith generates the graph. The framework details are handled for you, though developers can read and edit the underlying spec.
The Smith runtime compiles each generated graph into an executable agent system, hosts a versioned endpoint, and streams traces. The studio is the design surface on top of it.
Yes. Select any agent to change its model, temperature, tools, and system prompt. Smith suggests a model per role; you can override it on any node.
Smith Studio is in early access. You can explore the full build-to-deploy loop in the live prototype now; real generation and runtime are rolling out in phases.
Describe the work, and watch Smith forge it into a deployed, traced multi-agent system.