Launch is a real, complete autonomous corporation — not a crippled trial. You bring zero AI expertise; the platform charters a full agent corporation built to do the day-to-day work, and you decide the business it runs. Higher plans don't unlock missing features — they grow the elastic frontier agentic workforce and its ceilings, from a starter corporation all the way to enterprise scale. Every plan ships the same security architecture — deny-by-default, signed supply chain, hash-chained audit trail, human-held kill-switch — with no safety feature held back for a higher tier. There is no card to enter today, and the plans below are a truthful mirror of the platform's in-tree billing catalog.
Questions? Email a human — a person replies: aipeteaipete@gmail.com
early accessFree during early access — this platform collects no payment and asks for no card. Paid tiers activate later through the platform's billing seam.
the model, at a glance
Read it in one look: every tier is a complete corporation with the same security baseline — what climbs is the size of the frontier agentic workforce and its ceilings, never a corp count and never a billed seat.
Caps are signals enforced fail-closed; an idle frontier agentic workforce scales to zero, so an idle corporation costs nothing. Exact per-tier numbers are in the cards and matrix below.
the plans
Every plan runs exactly 1 corporation, and every one of them is a complete corporation from the first tier up. What grows as you climb is the elastic frontier agentic workforce (worker replicas) and the concurrent-mission ceiling — more capability, more scale, more frontier intelligence — never a corp count, and never a billed seat. Pricing is a platform base fee plus metered usage; human admins are a membership allowance, not a per-seat charge. List prices are placeholder figures from the platform catalog; caps are signals, enforced fail-closed by the platform.
$0 /mo
free — base + usage · starter isolation tier · a real, full corporation to start
$49 /mo
base + metered usage · starter isolation tier · steady small-scale work
Most Popular
$349 /mo
base + metered usage · pro isolation tier · scales when work spikes, with SSO and an admin team
$1,299 /mo
base + metered usage · pro isolation tier · a larger frontier agentic workforce and admin group
custom · from $8,000 /mo
base + metered usage · enterprise isolation tier · the highest ceilings, isolated infrastructure, and your own keys
honest noteHow this works, plainly: pricing is a platform base fee plus metered usage (cost of goods × markup), and the listed fees are placeholder list prices mirrored from the platform catalog — authoritative pricing lives in the platform pricebook. Human admins are a membership allowance, not a billed per-seat charge; Overdrive and Foundry can add admins beyond the included allowance through an admin add-on whose price is a pricebook seam, not set on this page. The caps shown are plan signals — spending is enforced by the platform's fail-closed budget pre-flight (debit before each call; refuse at zero), never by this page — and an idle frontier agentic workforce scales to zero, so an idle corporation costs nothing.
capability by tier
The dials grow as you climb; the security baseline does not move. The bottom group is an intentional wall of identical checks across all five tiers — that redundancy is the point: safety is never an upsell.
| Capability | Launch | Studio | Overdrive | Foundry | Vanguard |
|---|---|---|---|---|---|
| Scale | |||||
| Corporations | 1 | 1 | 1 | 1 | 1 |
| Worker replicas | 2 | 5 | 25 | 75 | 250 |
| Concurrent missions | 1 | 3 | 15 | 40 | 150 |
| Monthly usage ceiling | $0 | $250 | $3,000 | $15,000 | custom |
| Base fee / month | $0 | $49 | $349 | $1,299 | from $8,000 |
| Access & admin | |||||
| Admin allowance | 1 | 1 | 5 | 10 | unlimited |
| Admin add-on (per admin) | — | — | ✓included | ✓included | — |
| SSO (SAML) | — | — | ✓included | ✓included | ✓included |
| SCIM provisioning | — | — | ✓included | ✓included | ✓included |
| Isolation & data | |||||
| Isolation tier | starter | starter | pro | pro | enterprise |
| Silo isolation per tenant | — | — | — | — | ✓included |
| BYOK / CMEK | — | — | — | — | ✓included |
| GDPR data export | — | ✓included | ✓included | ✓included | ✓included |
| Security baseline — identical at every tier | |||||
| Data-vs-instruction separation | ✓included | ✓included | ✓included | ✓included | ✓included |
| Human-rooted trust-tier boundary | ✓included | ✓included | ✓included | ✓included | ✓included |
| Safety as a disqualifying constraint | ✓included | ✓included | ✓included | ✓included | ✓included |
| Deny-by-default tools & egress | ✓included | ✓included | ✓included | ✓included | ✓included |
| Signed-skill supply chain | ✓included | ✓included | ✓included | ✓included | ✓included |
| Hash-chained audit trail | ✓included | ✓included | ✓included | ✓included | ✓included |
| Human-held kill-switch | ✓included | ✓included | ✓included | ✓included | ✓included |
| Fail-closed budget pre-flight | ✓included | ✓included | ✓included | ✓included | ✓included |
Numbers are placeholder list values mirrored from the platform billing catalog; caps are signals enforced fail-closed. No per-seat billing — admins are a membership allowance, and the add-on is priced per admin in the platform pricebook, not here.
trust & questions
Honest answers to the questions that matter. No certifications, customers, or metrics are claimed — only the architecture as it exists in the open repository.
Yes. Launch is a complete autonomous corporation, not a feature-gated demo — same security architecture, same agents, same guardrails as every paid tier. The higher plans grow the AI team and its ceilings; they do not switch on a corporation that Launch was missing. And to be plain about what we do and don't do: the platform provides the frontier execution — agents built to do the work — but you choose the business, the vertical, and the strategy. Those are real business decisions, and they're yours; we take the guesswork out of running it, not out of deciding what to build.
mechanismone corporation, fully chartered at every tier; you set the business, the platform executes it
No. The same security architecture applies to every tier from Free up. Deny-by-default tool access, a signed and provenance-gated supply chain, a hash-chained audit trail, and a human-held kill-switch are baseline — never an upsell. We do not sell safety as a feature.
mechanismone fail-closed policy plane mediates every tier; no tier bypasses it
No card during early access. Signing up collects no payment, and there is no billing path to accidentally trigger. Paid tiers activate later through the platform's billing seam, on terms stated plainly when that seam goes live.
mechanismdeny-by-default billing seam — off until explicitly configured at go-live
Per-tenant memory is encrypted, and every paid plan from Studio up includes GDPR data export. The enterprise isolation tier (Vanguard) adds silo isolation per tenant and BYOK / customer-managed keys (CMEK) so your keys stay yours. This static site itself makes zero external requests: no cookies, analytics, or trackers.
mechanismencrypted per-tenant memory; GDPR export; enterprise-tier silo isolation + BYOK/CMEK
No. There is no per-seat billing. You pay a platform base fee plus metered usage, and human admins are a membership allowance attached to your plan — one on the starter plans, more on the higher plans, unlimited on the enterprise tier. Overdrive and Foundry can add admins beyond the included allowance through an add-on, but the everyday price is base plus usage, not headcount.
mechanismbase fee + metered usage; admins are an allowance, not billed seats
No. They are placeholder list prices mirrored from the platform's in-tree billing catalog while early-access pricing is finalized. Authoritative pricing lives in the platform pricebook, and the caps shown are signals — the platform's fail-closed budget pre-flight is what actually stops spending at your ceiling.
mechanismcatalog mirror kept honest; pricebook is the source of truth
Self-serve isn't live yet — during early access we set up your full corporation with you, starting free. Book a 15-minute concierge setup to begin. Climb to a higher plan when you need a bigger frontier agentic workforce, more concurrent missions, more frontier intelligence, and a higher ceiling — up to enterprise. The security architecture never changes, and you always run one corporation.
Questions? Email a human — a person replies: aipeteaipete@gmail.com
The Agentic Corporation is in active development. This page mirrors the architecture and the billing catalog as they exist in the open repository — no customers, metrics, or certifications are claimed.