Pay for what runs. Nothing else.

Your company’s software (its database, warehouse, ontology, app, pipelines and agent) is billed on what actually runs: compute, storage, queries, app traffic, the company agent and AI evaluations. No seats, so every person and every agent can be added. Unit prices are published at launch.

Enterprise

Custom

For companies that want it built for them.

The whole platform, with nothing held back
A fixed-scope pilot, credited to your contract
Forward-deployed engineers for migrations and builds
Parity reports before every switch-over
Roles and permissions for people, API keys and agents
Support with agreed response times
Volume pricing on usage

Start a pilot

Self-serve · waitlist

Pay as you go

For teams who build with their own agents.

Everything: database, warehouse, ontology with roles, MCP and API, app hosting and the company agent
21 ready-made connections, plus any API with a key, OAuth 2.0 or client credentials
Connect any agent: Claude, ChatGPT, Cursor and more
Unlimited members, apps and agents
EU or US region
Usage billed monthly, by the meters below

No seats

Invite the whole company. Apps, agents and people cost nothing to add.

No feature gates

Every workspace gets the whole platform. Nothing is held back for a higher tier.

Pilot first

Our engineers build the first version on your data in two weeks, at a fixed price. Self-serve opens to the waitlist.

No idle cost

Nothing runs, nothing bills. No idle clusters, no minimum commitment.

What you’re billed on

Everything on the platform is billed on one of these meters. Unit prices are published at launch.

MeterWhat it coversUnitPrice

Pipeline computeIngestion, models and jobs running on managed computeper compute-hourAt launch

Warehouse storageData stored in your workspace warehouseper GB-monthAt launch

Warehouse queriesData processed by queries from pipelines, apps, agents and the CLIper TB processedAt launch

AppsTraffic to apps deployed from your workspaceper request and GB servedAt launch

DatabaseManaged Postgres under your ontology and appper compute-hour and GB-monthAt launch

AgentModel calls made by your company agent (new)per model usageAt launch

AI evaluationsDecision-model questions answered in your pipelinesper thousand evaluationsAt launch

Events (preview)Events received by your collection endpointsper million eventsAt launch

Questions

Is there a free tier?

You will be able to start without a contract and pay only for what you use. Launch details, including any starting credits, go to the waitlist first.

When will prices be published?

Unit prices are published at launch. The meters above are what you will be billed on; nothing else.

How is the company agent billed?

Its model calls are billed on the Agent meter, like every other part of the platform. There is no separate seat or licence for it.

How is enterprise priced?

It starts with a fixed-scope, fixed-price pilot on one solution, credited to your contract if you continue. After that, forward-deployed engineering is scoped with you per project. Platform usage is billed the same way as self-serve.

What happens to our data and code?

It lives in your workspace, in the region you chose, and everything your agent wrote is code in your workspace's git repository. If you leave, you take the repository and export your data.

Start with a pilot, or join the waitlist

Start a pilot

Our engineers build the first version on your data in two weeks, at a fixed price.

Or join the waitlist for self-serve:

Self-serve opens to the waitlist first.