Compare plans

BigQuery compute plus storage cost for each full calendar month. Every self-serve plan covers up to $50k/mo. The band is soft: we never cut off your data. Two months over $50k in a row and we ask you to move to Enterprise.

How many people can use the account. The trial has 1 seat.

How often we scan your projects on our own.

How many days of BigQuery job activity each scan looks back over. Dashboard charts go back the same number of days.

The ranked backlog of cost problems, with a dollar estimate and evidence on every item.

Learns your normal spend and storage pattern and flags the spikes that don't fit it. Every scan checks each day, and each block of time since the last scan.

Group spend by department or workload with your own rules, give each one a monthly budget, then filter the dashboard and usage views by cost center.

Check totals against your billing export so the number on screen matches the invoice.

Get notified when an anomaly or a new high-value finding shows up.

A forecast after each scan of your spend against your commitment and each cost center's monthly budget, plus a report by email with an Excel file: spend, top offenders, efficiency, anomalies and findings. Owners pick which parts each person gets. Starter and the trial pick weekly or monthly. Growth, Pro and Enterprise can also pick daily.

Recommendations for resizing BigQuery reservations and slot commitments. The 30-day trial includes this page. Starter and Growth do not. Reservation findings stay on the Findings page on every plan.

CSV download of the Findings page on every plan, and the dashboard's Excel report, sent to your email: the same file the report email attaches.

Pull spend, findings and storage data into your own tools with read-only API keys.

We read dbt details from the comments dbt puts on each BigQuery job. Nothing to set up.

Upload a dbt manifest, or let us sync dbt Cloud once a day.

See each dbt model with the cost of everything that feeds it.

Cost findings that use dbt context, like models that run often but are rarely read.

The AI DBA opens pull requests that fix dbt models. Nothing is ever merged for you.

A new fix pull request waits until one of the open ones is merged or closed.

How many GitHub repos the AI DBA can open pull requests in.

We trigger a dbt Cloud CI run on fix pull requests so you can check them before you merge. Each run uses your dbt Cloud job minutes.

One monthly allowance for Cognis AI chat, the AI query optimizer and Cognis AI DBA. Growth has 2.5x Starter's, Pro 5x. It resets each month on your billing date. The trial has no AI.

Ask about your spend in plain English, and get AI rewrites for expensive queries, checked by dry run when you allow it. Both use your AI usage.

A policy-gated AI DBA that applies a small set of safe, reversible fixes — owner approval by default, every change logged and almost all reversible. AI-written fixes use your AI usage.

The AI DBA opens pull requests for dbt and code fixes on its own branches. Nothing is ever merged for you.

A record of every AI DBA action: who approved it, what changed and when. Each event is kept for 12 months.

The scanner service account never requests the role that reads your table contents, on any plan. The only optional exception is a linked Cloud Billing export, a single dataset of cost rows.

A full log of logins, role changes, API keys, exports, approvals and AI actions — who, what and when.

Set a device limit per person (down to one device only) and sign people out after a set idle time.

A contracted response-time commitment from support.

A shared Slack channel with your team and a named account manager.

A recurring review of savings realized and what's next, with a written report.

A Data Processing Addendum and help completing your security questionnaire. SOC 2 is in progress, not yet certified.

Custom contract terms and invoicing instead of self-serve billing.