Everything Layerbase does, on one page.

Provision any of 18 database engines from one dashboard, one account, and one monthly bill.

• Relational, document, key-value, analytics, search, vector, time series, graph, and ledger engines
• A per-database hostname and connection string issued at create time
• Deep-linkable create flow, so a link can preselect the engine for a teammate

Read the docs

Choose the major version you want at create time instead of taking whatever the platform pinned this month.

• Prerelease versions are labelled as prerelease before you pick one
• Older versions stay running when a version stops being offered for new databases

Point Layerbase at a database you already run somewhere else and it copies the data over for you.

• Connection-string import, a preflight check, and a progress readout on the migration
• Guides for the platforms people leave most often, including Postgres, MySQL, Redis, and document sources
• Migrate into a brand new database, or into an existing empty or running one

Read the docs

Move an existing database onto a compatible engine without rebuilding it by hand.

• MySQL to MariaDB keeps the same address, with about a minute of downtime
• Redis and Valkey convert in either direction
• SQLite converts up to libSQL when you want remote access and auth

Read the docs

Create a database with a lifetime attached so a test run cleans up after itself.

• Set a time to live between 1 and 72 hours at create time
• Counts against your database quota while it lives, then removes itself

Read the docs

Create databases from a script, a CI job, or an agent instead of clicking through the dashboard.

• A monthly ceiling on API-created databases per plan, so a runaway loop cannot drain an account
• Databases you create in the dashboard never count against that ceiling

Read the docs

An idle database hibernates instead of billing you for nothing, and comes back when a client connects.

• Free databases idle out after 15 minutes, or 30 minutes on the slower-waking MySQL and MariaDB
• Waking is automatic on connect for most engines; MySQL and MariaDB take longer to come back, so start them first if a cold connect would time out
• Stop and start a database yourself whenever you want, and a stopped database stays stopped

Read the docs

Pin a database so it never sleeps, funded by the always-on memory pool that comes with your plan.

• The pool is a capacity budget: pin as many databases as it fits
• Pro can add pool blocks at $10 per month for more room

Read the docs

A free database that has sat unused for weeks is archived rather than destroyed, and warned about first.

• Archiving applies to the Free tier only, after 14 hibernated days (7 for a database never used)
• Warning emails go out before anything happens, and again as the window closes
• Upgrading to a paid plan takes a database out of the archive pipeline entirely

Read the docs

Bring an archived database back with a single button, and the clock resets from there.

• Restore takes about 10 to 30 seconds and returns the same data
• A database left archived past 90 days has its live resources reclaimed, and its most recent backup is KEPT and stays restorable

Read the docs

Drop to a smaller plan and anything over the new limits is locked, not thrown away.

• A grace window with warning emails before anything locks
• A locked database can still be downloaded, and unlocks the moment you upgrade or free up room

Read the docs

Keep a database tidy over its life without breaking the connection string your app already uses.

• Rename the display name, or the real database name, without changing where clients connect
• Tag a database as production, staging, development, or preview
• Apply engine image updates yourself when one is available, with a short restart

Every destructive action backs up before it runs, and refuses to proceed if that backup cannot be secured.

• Deleting a database requires typing its name, on every plan including Free
• The pre-teardown backup is kept and remains downloadable

Automatic backups on a rolling retention window, with no job for you to configure.

• Backups are stored off the server that runs your database
• Dedicated servers can also choose their own cadence, down to hourly
• An account-wide view shows the latest backup for every database at once

Read the docs

Take a named snapshot before a risky migration and keep it exactly as long as you want.

• Name and rename a backup so you know what it was taken for
• Replace the oldest slot when they are all full, with a confirmation first
• Your last remaining copy of a database can never be deleted

Read the docs

Pull a backup down to your own machine, including from a locked or archived database.

• No plan gate: a Free account can always take its data out
• Works while a database is locked, archived, or offloaded

Read the docs

Roll a database back to any backup you still hold, from the dashboard, in one flow.

• Restores are exercised on a schedule against real backups, not assumed to work
• Restore into the same database, or use a backup to seed a new one

Read the docs

Each backup carries a checksum and a verification state, so you know it is intact before you need it.

Read the docs

Opt a branch into the same backup schedule as the database it came from.

• Off by default, because most branches are disposable
• Turn it on per branch when one is doing real work

Read the docs

Rewind a database to a moment before the bad migration, not just to the last backup.

• PostgreSQL and FerretDB, on Pro and above
• The database has to be pinned always-on, since the recovery log needs a running instance

Read the docs

Take a full export of a database at any time, in any state, with no plan gate on it.

Read the docs

Make a working copy of a real database with its real data, the way you branch code.

• Available on 16 of the 18 Cloud engines
• A branch does NOT consume one of your plan database slots
• Each branch gets its own connection string and its own credentials

Read the docs

Turn on instant branching for a database and new branches share storage instead of copying it.

• A per-database switch, so you opt in only where you branch often
• A few engines are addressed by port rather than hostname, and get a new connection string when you enable it

Read the docs

Throw away everything a branch did and pull fresh data down, keeping the same connection details.

• Host, port, and credentials survive the reset, so nothing has to be reconfigured

Read the docs

See which database a branch came from and what has branched off it, as a tree.

• Branch badges on the database list so a branch is never mistaken for a parent
• Deleting a parent offers to take its branches with it, and says so before it does

Read the docs

Mark a branch as preview or development and it sleeps aggressively, because it is meant to be thrown away.

Read the docs

One dialog with everything needed to connect: the string, the parts, and working code.

• Copy-ready snippets for the CLI, Node, Python, and HTTP clients
• A preview of the exact fields a desktop GUI client asks for
• Individual host, port, user, password, and database fields, masked until you reveal them

Read the docs

Use a pooled endpoint so a serverless function fleet does not exhaust your connection budget.

• Pooled and direct endpoints are both issued, and both are shown in the connect dialog
• Tools that need a real session, like some migration runners, use the direct endpoint

Read the docs

Every connection is encrypted, and your databases run in your own container on your own private network.

• TLS on every engine endpoint, with end-to-end passthrough on PostgreSQL
• A container and private network per customer, not a shared multi-tenant instance

Read the docs

Your database is addressed by a hostname that does not change when we move things around underneath it.

• Most engines are reached by hostname on a standard port for that engine
• A few engines are addressed by their own port instead, which the connect dialog spells out

Read the docs

Restrict a database to the addresses you name, on every plan, including Free.

• Per database, managed from its settings, effective immediately
• A helper script adds the address you are currently connecting from

Read the docs

Require a client certificate to connect, so a leaked password alone is not enough.

• PostgreSQL only, on Pro and above
• A certificate authority per database, with one-click rotation
• Certificates are issued alongside the password, never instead of it

Read the docs

Issue a key scoped to a single database, and rotate it without an outage.

• Shown once at creation, then stored hashed
• Rotation keeps the old key working for a grace period so deploys can catch up

Read the docs

Support cannot see your credentials unless you grant access, and the grant expires on its own.

• You choose the window when you open a grant, and can revoke it early
• Without a live grant, credentials are withheld from staff tooling

Learn more

Speak to a Layerbase database with the serverless drivers your framework already ships.

• An HTTP endpoint for Redis and Valkey that matches the popular REST client
• Serverless driver compatibility for PostgreSQL and for MySQL and MariaDB
• A direct TLS endpoint for runtimes that cannot negotiate a database-native handshake

Read the docs

Write and run SQL against a running database from the browser, with results in a real grid.

• Multi-statement runs, with a confirmation before anything destructive
• Query formatting and starter queries for the engine you are on

Read the docs

Query time-series data and see the result as a chart instead of a wall of rows.

Read the docs

Browse keys, inspect values, and run commands against Redis and Valkey without a local client.

• A key browser that pages through the keyspace safely

Read the docs

Run document queries against FerretDB with the shell syntax you already know.

Read the docs

Send requests to engines that speak HTTP without leaving the dashboard or building a curl line.

Read the docs

Try searches, tune index settings, and issue scoped keys for a search index from the dashboard.

• Edit index settings such as searchable and filterable attributes
• Create scoped keys so a client only sees the indexes you allow

Read the docs

Create accounts and post transfers against a double-entry ledger through a guided form.

Read the docs

The console keeps your place: tabs, history, saved snippets, and keyboard-first navigation.

• Multiple query tabs, kept in your browser between visits
• Recent query history and saved snippets, stored locally to your browser
• A command palette and keyboard shortcuts for running, formatting, and switching tabs
• Schema browser and autocompletion for tables and columns on SQL engines
• A results grid with a row detail panel, and export to CSV, TSV, JSON, JSONL, Markdown, or INSERT statements
• A hibernated database wakes automatically when you run a query against it

Read the docs

Drop a file into the console and load it into a table without installing anything.

• CSV, TSV, and SQL uploads, with larger files handled through a direct upload
• Run a whole SQL file against the database in one go
• DuckDB can also read JSON, newline-delimited JSON, Parquet, and remote URLs

Read the docs

Run a query over plain HTTPS from an edge runtime that cannot open a database socket.

• Available on 17 of the 18 Cloud engines
• Authenticated with a database API key, rate limited per account

Read the docs

When a query fails, ask for an explanation and a suggested fix without leaving the console.

• Reads the failing statement and the error the engine returned, and proposes a correction
• You review and run the suggestion yourself; nothing is applied for you

Read the docs

Put databases in a shared workspace so the people you work with can reach them on one bill.

• One member seat is included; extra seats are billed per member per month
• Owner and member, with no role hierarchy to configure

Read the docs

Invite a teammate by email and they join through a link that expires on its own.

• Invitations expire after a week if they go unused
• A notification in the dashboard when an invitation is waiting for you

Read the docs

Hand an existing database over to a team without recreating it or reissuing its connection string.

• Shared databases are labelled so it is obvious whose they are
• Removing a member removes their access without rotating credentials for everyone else

Read the docs

Rename a team, transfer ownership, or wind one down when a project ends.

Read the docs

One price per month per plan: Free at $0, Solo at $5, Pro at $15. No usage meter anywhere.

• Free includes 2 databases and 5 GB, with no card required
• Solo includes 2 databases and 10 GB
• Pro includes 10 databases and 25 GB

Learn more

Try Pro for 7 days before you commit to it.

• A card is required, and one trial per customer
• End the trial early from the dashboard whenever you want

Read the docs

Add capacity to a Pro account in $10 steps instead of jumping to a bigger tier.

• Each block adds 1 GB of always-on memory, 1 vCPU, and 25 GB of storage
• Up to 20 blocks on one account

Read the docs

Buy a private single-tenant server at a flat monthly price and put as many databases on it as it holds.

• Self-serve presets, provisioned in about 15 to 30 minutes
• Choose which engines are warmed on your own box
• Prepaid terms are available for a longer commitment
• Close a server when you are done, after a fit check on what is still running on it

Read the docs

Generous rolling allowances rather than a bandwidth meter, so a busy week never produces a surprise invoice.

• Rolling 24-hour, 7-day, and 30-day windows, summed across all your databases
• A modest overage gets a grace window and an email, not a shutdown
• Dedicated servers are unmetered: the hardware is yours

Learn more

Upgrade, downgrade, cancel, or resume from the dashboard, with the maths shown before you confirm.

• Switches are prorated, and a preview shows what changes on your account first
• A billing portal for invoices, cards, and receipts
• An activity feed of what changed on your subscription and when

Read the docs

Credits on your account are applied to your bill automatically, and shown on the plan page.

Read the docs

Connect a project and get the database environment variables set for you, per environment.

• Environment variables written and kept in sync, including a pooled and a direct variant
• Optionally reset the staging database on each deploy, so previews start clean
• Cancel a queued reset and redeploy without unwiring the integration

Read the docs

Install the Layerbase GitHub App so branch activity in your repository can drive your databases.

• Read-only repository access: Layerbase never writes to your code
• A guided wizard for picking the repository and the database it maps to

Read the docs

Give every git branch its own database branch, created when the branch appears and cleaned up when it goes.

• Works from either the GitHub App or the Vercel integration
• Preview environments get their own data instead of sharing one staging database

Read the docs

See at a glance whether a connected project is still wired up, and what the last automation run did.

• Connection status for each linked project
• Recent automation errors surfaced where you connected the project, with what to fix

Read the docs

Run a hosted message queue on top of a Valkey database, included at no extra charge on every plan.

• Valkey only, and free on every plan including Free
• Scheduled delivery, retries with backoff, and a dead-letter queue
• Compatible with a widely used queue client, so existing code often works unchanged

Read the docs

Create, list, query, back up, and delete databases over a documented HTTP API.

• An endpoint that reports your plan and how much of your create allowance is left
• The same operations the dashboard uses, so nothing is dashboard-only

Read the docs

Hold an account-scoped key for tooling that manages more than one database.

• Account-scoped and database-scoped keys, listed and revocable in one place
• Shown once when created, then stored hashed

Read the docs

Manage local and cloud databases from your terminal with a single npm-installed command.

• One install puts two commands on your path, layerbase and the shorter lbase
• Every local command runs with no account at all; sign in only to reach your cloud databases
• A drop-in for SpinDB, which continues as the open-source engine underneath: the same commands, flags, and exit codes
• Log in from the browser once, or set an API key so CI and coding agents authenticate with no browser and no token to expire
• With a key the CLI talks straight to the REST API, under the same quota and rate limits the dashboard enforces

Read the docs

Run any of the 21 engines in the suite on your own machine as native processes, with no container runtime and no account.

• Real database binaries run as ordinary processes, so there is no daemon to keep alive and no virtual machine overhead
• Each database gets its own data directory, so two versions of the same engine never collide
• The same commands on macOS, Linux, and Windows, on ARM and x64
• Stopping a database keeps its data; only an explicit delete removes it

Read the docs

Fork a local database copy-on-write and let it follow your git branch, without your connection string ever changing.

• Instant and near-zero-space on APFS, Btrfs, XFS with reflinks, and ZFS, with a full-copy fallback everywhere else
• Check out a git branch and the matching database swaps in while DATABASE_URL stays exactly the same
• A running source is stopped, snapshotted, and restarted for you, and the fork starts on its own port
• List, reset, and prune branches, and the JSON output says whether a fork was a reflink or a copy

Read the docs

Turn a local database, a SQLite or DuckDB file, or a Postgres dump into a hosted database in one command.

• Creates the matching cloud database, loads the data, and prints the connection string and the dashboard link
• Files are identified by their header rather than their extension, and anything unsupported is refused before a cloud database exists
• Can write DATABASE_URL into your local env file, rewriting only that line and leaving the rest alone
• A failed load never deletes anything on its own: it says what happened and prints the retry and delete commands

Read the docs

Open a client against a database without the password ever becoming a command-line argument.

• Credentials reach the client through an environment variable or a private temporary file that is deleted on exit, so nothing lands in shell history, in the process list, or in scrollback
• JSON output on the commands scripts read, so a pipeline never has to parse text meant for a human
• Documented exit codes for auth failure, a suspended account, a capacity conflict, and a quota limit
• Anything that mutates refuses to run non-interactively without an explicit confirmation flag

Read the docs

Manage a database you host somewhere else through the same commands, and pull a copy of it down to your machine.

• Auto-detects Neon, Supabase, PlanetScale, Upstash, Railway, and Aiven from the connection string you paste in
• A pull takes an automatic backup before it replaces anything
• Pull from a URL or from an environment variable, rename on the way in, or do a dry run first
• A parallel fast path for large PostgreSQL pulls, with the option to skip bulk tables

Read the docs

Package a local database for a container platform, even though running it never needed one.

• Generates a Dockerfile, a compose file, an entrypoint script, TLS certificates, and credentials
• Ready to build and run anywhere containers run, with no further hand-editing

Read the docs

Run and manage databases on your own machine, across all 21 engines in the suite.

• Free for personal use on macOS, Windows, and Linux
• macOS: a universal build for Intel and Apple Silicon, Developer ID signed and notarized, so it opens with no warning and no workaround
• Windows: fully working, though the builds are not yet code-signed, so SmartScreen may warn on first launch (choose More info, then Run anyway)
• Linux: .deb and .rpm packages
• Includes engines that licensing keeps out of Cloud

Read the docs

Machine-readable manifests so a coding agent can use Layerbase correctly without you pasting docs.

• A skill file and an agent manifest generated from the live product, not written by hand
• A curated site map for agents at /llms.txt and its full expansion

Read the docs

Scaffold authentication tables and wiring code against your own SQL database instead of renting a login service.

• A setup wizard that writes the schema and hands you the framework code
• Copy-ready snippets for several popular auth libraries and frameworks
• Social sign-in setup guidance for the providers you want to support
• This scaffolds auth in YOUR database; it is not a hosted login service

Read the docs

Browse and manage the user rows your auth library created, from the dashboard.

• How much you can do depends on the auth library whose schema is in the database
• A dedicated Auth section that lists every auth-enabled database in one place

Read the docs

A written guide for connecting to each engine, with the driver setup that engine actually needs.

Read the docs

One page showing storage, memory, bandwidth, and database count against your plan.

Read the docs

See what one database is using, and how much of the always-on pool it is holding.

• A pool meter showing what is pinned and what room is left
• Live status refresh, so a waking database updates without a page reload

Read the docs

Get told in the dashboard and by email as you approach a plan limit, before anything is enforced.

• Banners for storage, memory, and bandwidth pressure
• Lifecycle emails ahead of archiving, offloading, or locking anything

Read the docs

Sort, filter, and switch views over your databases, with an overview of what needs attention.

• Card and table views, with search across names and engines
• A home overview summarising counts, state, and anything that needs a decision

A public status page for checking whether a problem is yours or ours.

Learn more

Sign in with an existing identity provider, with no separate password to manage.

• Both providers resolve to one account when they share an email address

Change the email on your account, pick a theme, and manage what you hear from us.

• Email changes carry through to your cloud account, so nothing forks
• Light and dark themes across the dashboard and the site

Close your account from the dashboard, with a full erasure option if you want everything gone.

• A standard delete keeps your backups so support can reinstate the account if you change your mind
• A full erasure removes retained copies as well, and is confirmed explicitly

Read the docs

Targeted notices in the dashboard about the things that affect your plan and your engines.

Report a problem from the page you are on, and the report carries the context we need with it.

Learn more

Ask a question and get an answer from a person, alongside a searchable FAQ for the routine ones.

Learn more