Reading transactions is how your app or agent pulls the money that has already moved through an account, straight from the Platform API. Card activity and external bank transactions both come back from the same API, so an agent can read the full picture of what happened without stitching separate sources together.

This is a developer feature. If you're building an integration or connecting an agent that needs to reconcile, categorize, or report on real activity, these are the endpoints it reads from.

There are two ways to keep up with transaction data, and they work together. Webhooks push you a small notification the moment something happens, so your app knows there's something new. Reading from the API is the other half: your app or agent calls the Platform API to fetch the actual records, either in response to a webhook or whenever it needs the current state.

You can read card transactions and external transactions this way. Card transactions cover the activity on your Sequence cards. External transactions cover the movement on accounts you've connected from outside Sequence, which we pull from the bank on our own schedule. Reading both from one API means your agent sees the whole ledger in the same place.

Once an agent can read the ledger, it can do the work that sits on top of it. Common examples:

Yes. Reading is read-only, so these endpoints return information and never move money. Your app authenticates with a scoped API key, and an agent connected over MCP can only read what your account allows. Nothing leaves your accounts from a read call, and any money movement an agent starts still waits for your approval, exactly as it does everywhere else.

You call the Platform API with your API key. The exact endpoints, the fields each transaction returns, and example requests live in the developer docs, which stay current with every release.

Read the transactions endpoints in the developer docs