docs
briven is postgresql 18 with pgvector already switched on. anything that works with postgres works here, so most of what you already know is the documentation — these pages cover the parts that are ours.
the shortest useful example
Two things get you a row: a database, and a key. Create the database in the dashboard, create an API key on the api keys page, and then this works from anywhere with an internet connection — no driver, no connection to hold open.
bash
curl -sS https://briven.tech/api/v1/sql \
-H "Authorization: Bearer $BRIVEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{"query":"select $1::text as hello","params":["world"],"database":"notes"}'response
{
"command": "SELECT",
"rowCount": 1,
"rows": [{ "hello": "world" }],
"fields": [
{ "name": "hello", "dataTypeID": 25, "format": "text" }
]
}If you would rather use a real Postgres connection, take the connection string instead and point psql, Prisma or Drizzle at it — connecting covers what the pooler changes and what it does not.
when the caller is a program
A nightly job, a deploy step or your own application has no browser and no password to type. It carries an API key and talks over HTTPS instead. sql over http is the endpoint that runs a statement and hands back rows, and it is the one most programs need. It caps every request at thirty seconds, ten thousand rows, four megabytes and a hundred parameters, and those caps apply to every key on every plan.
The management api is the other half: with an organisation-wide key you can list your databases, fetch one and list its branches. Creating and deleting databases over that API is switched off for every organisation by default and there is no setting you can change — ask us to open it for your account.
There is also a cli, documented for completeness. It is not published; do not install @briven/cli expecting this tool — that package is an earlier product. The HTTP helper @briven/serverless is on npm; the endpoints above are still the way in if you would rather not add a package.
everything in here
start here
interfaces
- sql over httpone statement, one POST, rows back — with @briven/serverless, or any client that speaks the protocol.
- management apicreate, list and delete databases; branch them; rotate a credential.
- clithe same things from a terminal, without pasting keys about.
- ai agents (mcp)point claude or cursor at a database — read-only by default, and nothing that can delete.
features
- branchinga copy of a database to break, and what it costs to take one.
- history and undothe change log inside your database, undo, and restore points.
- vector searchpgvector is already on. the table, the index and the query to write.
- functionsyour typescript next to your database. no outbound calls. off by default.
what briven does not do
briven does not have a hosted login system for your own end users, logical replication or a schema-diff tool. functions exist — documented here — but they cannot make outbound calls, and the flag is off until we switch it on for your account. API keys, history and metrics are managed in the dashboard and are not on the management API. And usage billing is not switched on — every usage rate in the product is zero and no payment provider is connected.
Something missing that you needed? Tell us. These pages are written from the code that serves them, so if one of them is wrong it is a bug and it gets fixed like one.