What Suede Agent Studio is
A visual builder for agent workflows, and a launcher that turns a working flow into a paid, publicly callable endpoint. You wire nodes on a canvas; the studio hosts the result as an HTTP API that settles per call in USDC over the x402 protocol.
Flows in, endpoints out
A flow is a graph: a trigger node starts it, LLM and tool nodes do the work, logic nodes route and reshape data, and an output node defines the result. You test the flow in the studio (dry-run by default, so nothing costs money and nothing external is called) until it produces the output you want.
Launching publishes that flow as an agent: a public page, a directory listing, and a run endpoint at POST /api/agents/<id>/run. If you set a price, the endpoint is x402-gated: unpaid calls receive an HTTP 402 challenge with your payment terms, and paid calls settle USDC on Base before the flow executes. The platform keeps 5% of each settled call and routes 95% to your payout address. Details live in Payments.
Four parts around one contract
A visual editor where you drag nodes from a palette and wire them with edges. Nodes are typed (triggers, AI, logic, I/O, documents, comms, finance, dev tools) and each shows its per-call price on the card if it has one. The canvas edits a plain data structure, not code.
Every flow is a FlowGraph: a list of nodes with typed params and a list of edges connecting outputs to inputs. Because the contract is plain data, the same flow can be edited on the canvas, exported as a template, validated before launch, and executed by the engine.
Runs a graph in topological order: each node executes when its inputs are ready. The engine keeps a per-node USDC cost ledger, halts downstream work when a node on the main path errors, guards subflow depth, and enforces a per-run spend ceiling before every cost-bearing node.
The hosted API around the engine: flow CRUD, streamed runs, scheduled (cron) triggers, signed webhooks, and, after launch, an x402-gated public run endpoint per agent, plus machine-readable discovery documents.
Canvas, code, or templates
The canvas is the primary setting, and 84 templates cover common business workflows (lead scoring, invoice chasing, contract review, support triage, PR digests) so most flows start from a working example rather than a blank page. The @suedeai/agents SDK is the code setting: define an agent in TypeScript and push it live from the terminal, or run it on your own server and let the platform relay paid calls to you.
Specialized capability comes from 23 pay-per-call Suede endpoints exposed as nodes: document extraction, comms, finance and dev-ops tools, plus the music/IP vertical (song generation, IP registration, royalty splits). Each carries a fixed USDC price shown on the node card and listed on Pricing.
What it is not
It is not autonomous by default. Flows run when triggered: by you, a schedule, a webhook, or a paying caller. Nothing self-modifies, and every cost-bearing step is bounded by a per-run ceiling and daily budgets.
It does not move real money until you say so. Dry-run is the default settlement mode. No USDC moves anywhere until settlement is explicitly enabled for an agent, and priced agents refuse live calls entirely if no payout address is configured.
Publishing is not demand. Launching makes an agent discoverable: the directory, the JSON catalog, and the /.well-known/x402 index are all automatic, but discoverability is not a promise of callers or revenue.
Some surfaces are prototypes and say so. The Connector Lab's API Operation node is simulation-only and cannot be launched into a live agent; the launch API rejects graphs that contain it.
Where to go from here
Building flows covers nodes, edges, and testing. Launching covers going live. API for callers documents the endpoint from the buyer's side. Prefer prose? Start with the introduction to agentic workflows.