You pay for events and reads.
The price depends on how much you send, how much you read, and whether it runs in our cloud or yours. Start free; a quote takes one call.
Free tier
Prove it works on your own events before anyone talks about money.
- Up to 5 million events a month
- 1GB of attribute memory
- Unlimited attribute definitions, any entity key
- Same engine as the paid plans, shared region
- Snowplow Assistant and MCP server
- Docs and community support
Snowplow-hosted
We run the collection, the streaming engine and the serving layer. You write definitions and read them.
- Fastest to production. Nothing to provision
- Dedicated region, serving availability target
- Streaming and batch engines from one definition
- Events land in your warehouse
- Support with a response target and a shared channel
Your cloud
The same engine deployed into your own account, for teams whose data cannot leave it.
- Data stays inside your own account
- Your region choices, your network policy
- SSO, SCIM and audit log export
- DPAs, security review, procurement support
- Named solutions architect for the rollout
What you are quoted on: events in, reads out.
Events processed
Counted once on the way in. Failed validation still counts; we would rather you saw it than hid it.
Attribute reads
One API call to one service is one read, whether it returns three attributes or thirty.
Attribute memory
The free tier serves your live attributes from 1GB of memory. At the cap the oldest attributes are evicted first. Paid deployments are sized on volume, not memory.
Frequently asked questions.
Why is there no price on this page?
Because volume and deployment move it too much to publish one list. Start on the free tier without talking to anyone; get a quote when you know what you need.
What will you need from us to quote?
Rough monthly events, rough attribute reads, and whether it runs in our cloud or yours. Three numbers, one call. Terms are sized on the year, not your worst week.
Is the free tier the real product?
Yes, on a shared region with caps on events and attribute memory. Same engine, same latency profile, same SDKs.
What happens when the free tier's attribute memory fills up?
The oldest attributes are evicted first, so the newest keep serving. If you are filling 1GB, you have outgrown the free tier and it is time to get a quote.
Can we move from your cloud to ours later?
Definitions are declarative and portable, so the migration is a redeployment rather than a rewrite.
Try it before you price it.
The real engine, free, on your own events. Define one attribute, read it from your product, then work out what volume you actually need.