Pricing

One plan.
Priced per engagement.

There are no tiers to compare, because the part that varies is your systems. A one-time setup connects them and authors your first agents; a monthly subscription keeps them running and adds more over the contract.

Contract

Standard

Let's talk

One plan, priced per engagement.
A one-time setup, then a monthly subscription. Everything below is included.

Book a demo

Not sure yet? Book 15 minutes with us — bring the list of software you run.

How We Think About Pricing

01
One plan, and everything is in it. Agents, tasks, schedules, connectors, isolated computers, live view, and the full transcript. Nothing is held back for a higher tier, because there is no higher tier.
02
Price the engagement, not the seat. What differs between two customers is which systems have to be connected and how their work is done — not how many people log in.
03
Setup is real work, and we do it. Mapping your systems, building the connectors, authoring the agents, and training your team is the setup fee. It is not a signup form with a fee attached.
04
The subscription buys maintenance, not access. Your vendors change their APIs, credentials expire, and workflows accumulate. That is the ongoing work, and it is what the monthly fee covers.
05
Compute is passed through, not marked up. Every contract includes a compute and model allowance. Usage above it is billed per organization, at what it costs to run.

Frequently Asked Questions

What's an agent?

An agent is one configured worker: a set of instructions that encode how your company does a job, the connectors it may use, and the guardrails it works inside. You name your own agents — Dispatch, Invoice chaser, Intake. There is no assistant persona and no product-within-the-product.

What's a task?

A task is one execution — a conversation with an agent, or one unattended pass over a job. You watch the transcript and the agent's screen side by side, and you can interrupt and steer it mid-task.

What's a schedule?

A schedule puts an agent on a cadence: check trailer status in the yard system every morning, chase unpaid invoices every Monday, prepare charts the night before clinic. Schedules have gates and skip rules, and every execution keeps its own transcript.

What's a connector?

A connector is authenticated access to one of your systems — your fleet software, your mailbox, your EHR, your ERP. We build and configure them; you see the list, the status, and when each was last used. Most of the systems we connect to have no public API story, so this is engineering work rather than a checkbox. It is also the reason a vertical gets cheaper the second time we sell into it.

Which systems can you connect to?

Today the deepest coverage is trucking and logistics: fleet management, yard operations, and business email, with 32 tools live at our reference customer. Healthcare (EHR, scheduling, billing), professional services (CRM, billing, document stores), and field services (scheduling, equipment tracking, procurement) are the verticals we build into next. Bring the list of software you actually run to the demo. That list is what decides the setup estimate.

How long does setup take?

About a day and a half if you are in a vertical where we already have the connectors. Roughly one and a half to two weeks if your systems need new ones. Setup covers the assessment, the connector work, provisioning, credential setup, authoring the first agents, and one or two training sessions with your team.

What does it cost?

Per engagement, because the setup work is what varies. There are two parts: a one-time setup fee, and a monthly subscription that covers hosting, monitoring, connector maintenance, credential lifecycle, support, and the agents added over the contract. We quote both on the call, once we know which systems are involved.

Is there a free trial or a self-serve plan?

No. Every engagement starts with connector and configuration work on your systems, so there is nothing useful to hand you behind a signup form. The first step is a demo — we show you an agent working in software like yours, and you tell us what you would point one at.

Where do the agents run?

Each agent gets its own computer, with its own browser and desktop, so the blast radius of any one agent is bounded. By default those computers are ours, and the compute is billed to the contract. For regulated or air-gapped work — PHI, PII — the computers can run on your own infrastructure instead, and your data never leaves it.

Can we see what an agent actually did?

Yes, twice over. While a task is happening you can watch the agent's screen live and read the transcript as it streams. Afterwards the full record of every tool call and every action stays available for the term of the contract.

What if an agent does the wrong thing?

Four things stand between an agent and a mistake: guardrails written into its configuration, tool grants that scope what it is allowed to touch, approval gates on the actions you choose, and a live view you can interrupt at any point. Everything it did is in the transcript afterwards.

Do we have to change our software?

No. Agents work inside the systems you already run — they click, type, file, and follow up the way your team does. There is no rip-and-replace and no migration.

Tell us what you’d hand to an agent.

Half an hour, your list of systems, and one job you wish someone else did every morning.

Book a demo →