Connectors
How data comes into Canvasm — agent/API push today, native connectors rolling out.
Canvasm keeps a metric map honest by feeding it real numbers. There are two paths in, and they share one data model.
Two ways to bring data in
- Agent / API push — Current. Your agent (over MCP) or your pipeline (over the API) fetches data wherever it lives and pushes values into Canvasm. Available today and the most flexible path.
- Native connectors — Planned / rolling out. Canvasm connects directly to a source, fetches on a schedule, and normalizes the data for you. Rolling out source by source.
One canonical data model
Both paths validate against the same canonical schemas, so metric logic isn't reinvented per integration. The schema families at launch cover:
- Payments — payment, refund, payout, subscription, invoice
- Commerce — order, order line item, product, customer
- Analytics — event, page metric, funnel step, cohort
A canonical record carries where it came from, when it happened, and its typed fields, which is what gives every source-backed value its lineage.
Native connector status
Planned
Google Analytics 4 is the first native connector and is in setup. Additional sources (payments, commerce, product analytics) are planned and will light up as each connector ships. Until a connector is marked Current, use agent/API push for that source.
Connecting an account (how it will work)
When native connectors are available, you'll connect a source under Connected accounts in settings: authorize with the provider (OAuth) or enter an API key, and Canvasm stores those credentials securely on the server — the browser only ever sees a connection's status, never the tokens. Canvasm then fetches on a schedule and maps the results onto your metrics.