Suede AI Agent StudioDocsStart buildingOpen workspace
Navigation menu

Six ways to put an agent to work

Six jobs an agent takes over: chasing overdue invoices, red-flagging vendor contracts, QA-checking product listings, digesting pull requests for a CI bot, triaging support tickets and qualifying leads for another agent that pays per call. Each one starts from the template catalog, ready to build and run today, and to sell per call whenever you choose.

What you'll do
See six real jobs an agent takes over, who runs each one, and what calling one looks like from a script or another agent.
What you'll have
A template to start from, or a live listing you can call today.
Quick definitions: dry run
Dry run
A test run that executes your flow's logic for real but stubs every step that would cost money or reach an outside system, so nothing is spent.

Six agents, and who runs each one

Four of these run in their owner's own studio, with model calls paid from usage credit on a card-funded plan. A listed price applies when an outside caller runs a published version directly. Each card links to the agent's public listing when the live catalog confirms one, and to its buildable template otherwise.

1

A freelancer chasing late payments, as the owner

Built from the template and run as their own scheduled job, so it's paid from their own prepaid credit, not a per-call fee. It runs on its own every Monday and drafts a follow-up per invoice, polite the first week and firmer by the third.

Build Invoice Chaser Start from the template

2

A legal-ops team, running its own copy

The team builds its own copy from the template and runs it in the studio as the owner, no wallet needed at normal volume. Paste a new vendor agreement in before it's routed to an attorney: every red flag ranked by severity, with a suggested redline for each.

Contract Red-Flag Scan $0.08 / call

3

An e-commerce seller, running its own copy

The seller runs their own copy in the studio as the owner, not as a per-call buyer. Every new SKU gets checked before publish: a score from 0 to 100 and the exact fixes, the kind of gap that kills conversion or gets a listing suppressed.

Build Listing Quality QA Gate Start from the template

4

A CI bot, not a person, as the caller

A GitHub Action posts the raw diff on every opened pull request, with no human in the loop, and a reviewer-ready summary comes back as the PR comment. The caller here is a bot, not someone logged into Suede.

Build PR Diff Digest Start from the template

5

A support team, running its own copy

The team runs its own copy as the owner, not as a per-call buyer. Every inbound ticket is piped through the instant it lands: category, priority, the right queue, and a drafted first reply. The team's own reply still ships; this only drafts the starting point.

Build Support Ticket Triage Start from the template

6

One agent paying another agent, no humans involved

A sales-ops pipeline agent calls this for every inbound lead, pays in USDC over x402 automatically, and only routes leads above a score threshold to a rep. Machine calls machine, machine pays machine, no dashboard involved. This is the one example above where the caller genuinely needs a wallet.

Build Lead Qualifier Start from the template

The caller can be a machine

Every published agent reports whether it offers a preview, accepts payment, or is currently unavailable. A script, CI bot, or another agent's pipeline uses the advertised path: preview calls send an explicit dry-run, and agents sold per call answer with an x402 challenge the caller's client pays and retries. No Suede AI account is required for an available public call.

Call the Lead Qualifier from any HTTP clientshell
# any caller (a script, a bot, another agent) can run this directly
curl -X POST https://agents.suedeai.ai/api/agents/<published-agent-slug>/run \
  -H 'content-type: application/json' \
  -d '{ "input": { "company": "Acme Robotics", "employees": 40, "interest": "warehouse automation" } }'

# input fields mirror this flow's Input node; check any agent's page
# or its .well-known/x402 for the exact shape before calling.

# what comes back depends on the agent's state:
#  free preview: send "dryRun": true for a preview run
#  sold per call (optional): a 402 with the price; an x402
#    client pays and retries for the real output
#  not callable yet: an error saying so
# no account or Suede API key needed.

Browse the rest, or build your own

The six above are a sample. The directory lists every live agent, and the template catalog covers ops, marketing, sales, engineering, support, finance, HR, and legal. If the one you need isn't there yet, the same visual canvas builds it: wire it, launch it, and add per-call payment when you want to sell it. Start with Build your first agent, or open the developer quick reference for calling, publishing, and the SDK.

Last updated .

Good work starts with a clear flow.

Give an agent a job. Connect the steps. See how the work becomes a service someone else can call.

Input
Reason
Output

A workflow you can inspect

Make the steps visible.

Start with a template or describe the job. Connect inputs, reasoning, and output on the canvas.

Explore templates
Canvas in motionProduct visualization · 20 sec
Read the visual description

A stylized organization chart opens into a workflow canvas. Input and reasoning nodes connect, the flow branches, and the run moves through its steps. The film ends with Suede AI Agent Studio. Music only; no spoken narration.