AppSite Agents

Human and AI agents on one roster.

Everyone who does the work — the people on your team and the AI agents that help them — on one roster, with the rules that route work to each, a run button, and a log of everything that happened.

Who it’s for

Anyone who hands out work.

Owners

See who does what — people and AI.

One roster holds your team and your AI agents side by side, with what each may touch and what each did.

One rosterHuman + AIActivity
Team leads

Route work by rule, not by memory.

Assignment rules send each kind of work to the right agent — a new lead here, a failed scan there — and the log shows it landed.

Assignment rulesTriggersLog
Developers

Register an agent once, run it anywhere.

RegisterRun nowYour keys
The roster

Work that knows who owns it.

Agents is the registry every AI tool on the platform reports to, and the roster your people are on — so an assignment always has a name against it.

  • One roster — people and AI agents, edited inline, with what each may touch.
  • Assignment rules — each trigger routes its work to the right agent.
  • Run now — any agent, on demand, from its row.
  • Activity log — every run and every assignment, afterwards.
  • AI configuration — AppSite’s provider, or your own keys.

From a task to a done task in three steps

Register, route, run.

Step 1 · Register

Put everyone on the roster.

Your people and your AI agents, each with what they may do.

Roster
Step 2 · Route

Write the rule once.

Which trigger goes to which agent, and how — the log proves it went.

Rules
Step 3 · Run

Run it now, or let the rule.

Every run lands in the activity log with what it did.

Run nowLog

Agents on your own Claude plan

The appsite runner, from the bar: who answers, what it is working on, and what it cost.

Who answers

A Claude plan, named.

Each runner in the bar names the Claude account that answers its jobs, and says whether it is on, waiting for a sign-in, or off.

Account namedOn · waiting · off
The work

Fixes, repairs, scans, deploys, ships.

Every job a runner has queued, running or just finished, with its last line and how long it has run — each opens its Cyber Check record.

Five kinds of jobOpens the record
The cost

Charged, or on the plan.

Calls the API answered are charged; calls your Mac or Cloud Shell answered on the Claude plan cost nothing — and the bar shows what they would have cost on the API.

US$ chargedAPI-equivalent
Start

Agents comes with every plan.

Put your first agent to work today.

Register it, write one rule, press run — and read what it did in the log.