Management client

The management client is the SDK’s way of doing what Studio does, from a script. Use it for custom tooling, seeding and one-off migrations. For a whole backend kept in version control, use infrastructure as code instead.

It needs a secret key and must only run on a server or a developer machine. Never ship it to a browser.

Get started

import { createClient } from "@pawabase/client";
import { manage } from "@pawabase/client/manage";

const m = manage(
  createClient({ url: process.env.PAWABASE_URL!, apiKey: process.env.PAWABASE_SECRET_KEY!, environment: "development" }),
);

const tables = await m.resources.list();

import os
from pawabase import Pawabase

pb = Pawabase(os.environ["PAWABASE_URL"], os.environ["PAWABASE_SECRET_KEY"], environment="development")

tables = pb.manage.resources.list()

To act on another environment with the same key: manage(pawabase, "staging") in TypeScript, pb.manage.environment("staging") in Python.

Services

Each is one property on the client, matching a Studio page:

Service (TS / Python)	Studio page	Does
`resources`	Data	Tables: create, replace, remove, records, import and export, migrations, indexes
`schema`	Data → History	Every table against its definition; unmapped tables
`policies`	Access → Policies	Named rules
`schemas`	Data → Schemas	Reusable shapes
`routes`, `flows`	Logic	Routes; flows (including `run`, `validate`, `runs`)
`schedules`, `subscriptions`	Logic	Schedules (`run`); event subscriptions
`events`	Activity → Events	`emit`, `list`, `detail`, `graph`
`webhooks`, `inboundHooks`/`inbound_hooks`	Logic → Webhooks	Endpoints, `test`, deliveries, `redeliver`; inbound `url`
`jobs`	Activity → Jobs	`list`, `detail`, `retry`, `replay`, `failed`, `dismissFailed`, `queues`
`buckets`	Storage	Buckets, objects, share links
`mail`, `mailTemplates`/`mail_templates`	Mail	`config`, `log`, `sendTest`; templates
`realtimeChannels`/`realtime_channels`	Realtime → Channels	Channel rules
`keys`, `secrets`	API keys, Secrets	Keys (`create`, `rotate`, `revoke`); secrets (`set`, `rotate`, `rollback`)
`domains`	Settings	Custom domains
`observability`, `alarms`, `statusPage`/`status_page`, `health`	Activity, Status page	Requests, errors, logs; alarms; incidents; audit
`environments`, `releases`	Environments, API versions	`create`, `update`, `promote`, previews, blueprints; releases
`backups`	Backups	`download`, `restore`
`cache`	Data → Settings	`stats`, `invalidate`
`dashboards`	Dashboards	Read-only boards

Common recipes

Create a table and its table

await m.resources.create({
  name: "tasks",
  owner_field: "owner_id",
  fields: [
    { name: "title", type: "string", required: true },
    { name: "done", type: "boolean", default: false },
  ],
  operations: { list: { enabled: true, policy: "owner" }, create: { enabled: true, policy: "authenticated" } },
});
await m.resources.migrate("tasks");            // create the database table

pb.manage.resources.create({
    "name": "tasks",
    "owner_field": "owner_id",
    "fields": [{"name": "title", "type": "string", "required": True}, {"name": "done", "type": "boolean", "default": False}],
    "operations": {"list": {"enabled": True, "policy": "owner"}, "create": {"enabled": True, "policy": "authenticated"}},
})
pb.manage.resources.migrate("tasks")

await m.environments.update("production", { auth: { signup_enabled: true, require_email_verification: true } });

pb.manage.environments.update("production", auth={"signup_enabled": True, "require_email_verification": True})

update merges the keys you send. The keys are the ones Studio saves; see Configuring Auth.

Replay failed jobs after an outage

const preview = await m.jobs.replay({ status: "failed", queue: "mail", sinceMinutes: 120, dryRun: true });
await m.jobs.replay({ status: "failed", queue: "mail", sinceMinutes: 120 });

preview = pb.manage.jobs.replay(status="failed", queue="mail", since_minutes=120, dry_run=True)
pb.manage.jobs.replay(status="failed", queue="mail", since_minutes=120)

Read the request log

const failed = await m.observability.requests({ status: "403", limit: 50 });
const one = await m.observability.request(failed.data[0].request_id);

failed = pb.manage.observability.requests(status="403", limit=50)

Good practice

Prefer IaC for anything you want reviewed and repeated; use this client for scripts that act on data or operations.
Always do dryRun first for bulk operations (replay, redeliver_many, restore).
Keep the key in an environment variable; rotate it if it leaks.
Treat results as { data: [...] } for list calls.

Admin client

Manage users.

Infrastructure as code

Definitions as code.

Overview

Infrastructure as code

Define your backend

Logic and events

Files and mail

Ship it

Reference

SDK clients

Management client

Get started

Services

Common recipes

Create a table and its table

Replay failed jobs after an outage

Read the request log

Good practice

Admin client

Infrastructure as code

​Get started

​Services

​Common recipes

​Create a table and its table

​Change sign-in settings

​Replay failed jobs after an outage

​Read the request log

​Good practice

Admin client

Infrastructure as code

Get started

Services

Common recipes

Create a table and its table

Change sign-in settings

Replay failed jobs after an outage

Read the request log

Good practice