Commercial guide

Understand what a plan includes and when billing occurs.

Cloud plans combine a private workspace, dedicated compute, governed execution, and a defined amount of GPT-5.6 Luna research capacity.

Billing and capacity5 chapters
01

Included research hours

Cloud Team includes 125 research hours per month. Cloud Program includes 375 research hours per month. Each figure uses GPT-5.6 Luna as the common planning baseline so plans can be compared without requiring customers to understand internal token accounting.

  • Team: up to 5 members and one active investigation
  • Program: up to 25 members and multiple active investigations
  • Every plan keeps a visible capacity ledger
  • The plan-period ceiling is enforced before more work starts
02

The hosted launch model

Luna is both the listed capacity baseline and the only hosted launch model. ApolloBot rejects a configurable base URL, third-party router, or broader model allowlist so capacity, data handling, and provider receipts remain consistent.

i
Keep the model name in context

Pricing always presents the human-readable research-hour quantity first, followed by “GPT-5.6 Luna” in parentheses. Any future model option requires a new disclosed capacity and data contract.

03

Checkout and plan activation

Self-serve Team and Program purchases use Stripe Checkout. The customer is linked to an existing ApolloBot organization before Checkout begins, and plan entitlements activate only after a signed Stripe subscription event is reconciled.

  1. 01Sign in and choose a workspace
  2. 02Review the exact monthly plan
  3. 03Complete Stripe-hosted checkout
  4. 04Wait for the verified subscription event
  5. 05Open the workspace with the new entitlement
04

Metering and hard stops

The internal usage ledger is authoritative. Stripe receives only idempotent incremental overage events after the included allowance is exhausted. Reaching the research-hour or organization hard cap still checkpoints and stops the investigation.

  • No provider call before a worst-case hard-cap reservation
  • No entitlement from an unverified browser redirect
  • No duplicate subscription for an already active organization
  • Billing management remains restricted to an owner or billing role
05

Dedicated-compute capacity

A paid public plan includes one concurrent dedicated VM in its advertised Team or Program size. Provisioning requires a current, region-bound compute approval and remains separate from the approval to run the Luna research workflow.

  • Team: 4 vCPU, 16 GB RAM, 100 GB
  • Program: 8 vCPU, 32 GB RAM, 200 GB
  • Spot capacity can be interrupted
  • Every session auto-deletes after at most four hours
  • Additional or persistent capacity requires a separately priced agreement

See the lifecycle in motion.

Pose a question and watch the governed investigation demo advance stage by stage.

Run the demo