Start free. Pay for what your work uses.
Every resource is metered per unit and priced in the open, like your cloud bill. Pick a plan for included usage, then see exactly what each service costs.
Plans
Live prices are briefly unavailable. They will be back shortly; in the meantime, contact us for a quote.
How billing works
01
Metered per unit
Every resource your work uses is measured in its own unit: tokens, CPU-hours, GB-months, requests. Nothing is bundled or estimated.
02
Priced when it happens
Usage is priced at the moment it occurs, so a price change applies exactly from when it starts.
03
Credit first
Your plan's included usage is applied first. On Pro, usage beyond it is pay as you go; on Free, paid work pauses until next month.
04
One itemized invoice
Each month closes into one invoice with a line per resource, paid by card or on terms. Payments are processed by Stripe and appear as SEMOGRAM.
Prices by service
Pay-as-you-go prices per unit. Open a service for every resource, model and region, with worked examples.
AI
Model tokens for agents, authoring and predictions, priced per model.
Compute
Serverless compute for requests, agent turns, pipelines and predictions.
Sandboxes
Isolated sandboxes that run plugins, transforms and generated code.
Storage
Object storage for files, evidence, artifacts and datasets.
Data transfer
Requests and downloads made by dataset connectors.
Integrations
Calls to remote tools in connected apps.
Workflows
Durable workflows and queues that keep long-running work reliable.