Your plan is a monthly bundle of tokens at one flat price. Go past it and overage bills at $25 per million tokens at the end of the month, never above the ceiling you set. Change plans anytime, up or down. Most teams start by having our engineer build their first roles, $999 a month, then take them over. Pick a plan, drag the slider, and watch the bill.
01 · PICK A PLAN
WORKSHOP
$99/mo
5M tokens included
Light usage. A handful of agents, occasional tasks.
STUDIO
$249/mo
14M tokens included
A few agents running daily workflows. The usual starting point.
FOUNDRY
$499/mo
32M tokens included
Many agents running concurrent workflows across teams.
FACTORY
$999/mo
70M tokens included
Agents working around the clock across departments.
02 · START WITH US BUILDING IT?
RECOMMENDED START · A REAL PERSON ON OUR TEAM · ENGAGE OR DISENGAGE ANY TIME
Dedicated Engineer · $999/mo
Most teams start here. You get to real work sooner, and the bigger and more tangled the job, the more a built-for-you role pays off. Flat monthly rate while engaged, no hourly billing.
1We learn the job and build the role, and code any integration it needs.
2We test as we build, and none of that testing draws from your plan.
3Your team reviews it on real work; we iterate, then hand it off.
Studio includes 14M tokens for $249/mo. Add more with the slider, billed at $25 per million tokens at month’s end, up to the spending limit you set.
0
OVERAGE COST
+$0.00/mo
0 tokens
TOTAL MONTHLY
$1248.00/mo
plan + overage + engineer ($999/mo)
TOTAL TOKENS
14M
usable in any month
Drag the slider to add overage and the bill updates live. Start on the plan that fits today: you can switch plans anytime, up or down.
Studio plus overage is still your cheapest option: $249.00/mo for 14M tokens. Switch plans anytime, up or down.
Your Studio plan plus overage costs $249.00/mo. The Foundry plan includes 32M tokens for $499/mo: cheaper, with headroom. Switch plans anytime, up or down.
If you go past your bundle
$25 per million tokens
Same rate on every plan, prorated per token: 100K tokens over is $2.50. Billed once, at the end of the month.
You set the ceiling
Set a hard dollar limit in-app. When you hit it, no new work starts; tasks already running finish rather than stop half-done, so real usage can run past the ceiling by that in-flight work. No overnight five-figure bill. No surprise invoice on the 1st.
Prices exclude applicable taxes. Model exchange rates (what each model draws from your plan, explained in section 04 below) are listed in the model catalog and may change as models are updated. See the Terms of Service for full billing terms.
04 · MATCH THE MODEL TO THE WORK
Not every task needs your most capable model. Every model has an exchange rate on your plan’s tokens: powerful models draw more per task, lighter models draw less. Routine work, like sorting inbound email, lookups, and first drafts, runs well on a lighter model with a more favorable rate. Give the routine share to one and the same plan gets more done, with no change to your bill. Each role picks its own model, so you tune this role by role.
HOW MUCH OF YOUR WORK IS ROUTINE?
60%
60%
ON STUDIO · 14M TOKENS A MONTH
Everything on our default model
~4,200 AI tasks/mo
Right-sized mix≈2.3× the work
~9,600 AI tasks/mo
Slide right to send some routine work to a lighter model and see how much further the same plan goes.
Same plan, same bill: the mix frees up ~5,400 extra AI tasks a month. Bought as overage, that capacity would cost about $676/mo. To get it, set routine roles to a lighter model in the role’s settings, or your engineer does it for you.
Assumes routine work runs on Claude Haiku (exchange rate 0.04×: 1,000 tokens of work draws 40 from your plan) and the rest on our default model (0.67×). The most powerful models have a less favorable rate and earn it on contracts and hard reasoning. Estimates, not quotes: see the model catalog for every model’s exchange rate.