A cloud for
small software.

Build the app your team needs, then deploy it with sign-in, permissions, data, and backups included.

Build the tool your team is missing.

A claims queue for six people. A firing log for two. A laptop tracker for four. Off-the-shelf software does not fit these jobs, and custom development has usually cost too much.

Read YC’s request for a cloud for small software

Agents can build these tools quickly. Pylon handles the sign-in, permissions, data, deployment, and recovery work around them.

Build and deploy in one workflow.

The same Pylon app runs in the preview and in production. The agent can inspect its source, schema, permissions, tests, logs, and deployment state whenever it needs to make a change.

Build a small app Get the agent skill

Building claim-intake

01 · Describe

you“a claim intake form for our carriers — they file with photos, ops approves”

✓schema Claim, Carrier

✓policy carriers read only their own claims

✓page /claims — form, uploads, queue

02 · Preview

●preview live while the agent edits beside it

03 · Verify

✓policy test 4 allow · 3 deny — as intended

✓verify routes, assets, health — pass

04 · Publish

you“looks right — publish it”

live at claim-intake.pyln.dev

In Pylon Cloud

Build in an isolated workspace, use the live preview, and publish through Pylon’s deployment pipeline.

With your coding agent

Use Claude Code, Codex, Cursor, or your editor. The same project structure and diagnostics work locally.

The cloud parts are already connected.

Pylon includes sign-in, permissions, data, file storage, logs, backups, and rollback. You do not have to assemble a new stack for every tool.

Sign-in and permissions

Magic links, OAuth, OIDC, SAML SSO, expiring invites, organization roles, and row-level policies that deny access unless you allow it.

Database and files

SQLite without provisioning, Postgres when you need it, plus local or S3-compatible file storage.

Hosting and domains

Get a secure URL on the first deploy, plus a global edge network, custom domains, and idle sleep on Hobby.

Logs and recovery

Logs, metrics, audit history, volume snapshots, one-click restore, and rollback through deployment history.

Two hands pass an app card through a mechanical permission gate.

Permissions apply to every read and write

Control who can see and change data.

Invite the people who need the app, then define what each user or role can read or update. Pylon enforces those rules in the data layer.

schema/claims.ts

policy({
  entity: "Claim",
  allowRead:
    "auth.userId == data.filedBy
     || auth.role == 'ops'",
  allowUpdate: "auth.role == 'ops'",
});

claim-intake.pyln.dev · viewing as

Damaged pallet

Northline · PO 44821

$1,240

Missing carton

Cedar · PO 44795

$310

Crushed corner

Northline · PO 44768

$460

Late delivery credit

Cedar · PO 44741

$180

Same URL, same app — the runtime filters every row before it leaves the server.auth.role == 'ops' → 4 of 4 rows

Small software should outlive the afternoon.

Nobody should bet a decade of their team’s internal tools on a company that has been around for a year. The way out is part of the product, not a data export bolted on at the end.

The runtime is open source

One binary. Run it on your own box, your own container platform, or your own Postgres. Pylon Cloud runs the same binary you would.

The data is an ordinary database

A SQLite file, or a Postgres you can connect to directly. No proprietary storage layer in between — take a dump and it opens in any client.

The source stays yours

Export it, or connect a GitHub repository and keep every change there from the first deploy. Nothing built here is trapped here.

One plan for all your team’s tools.

Hobby gives you one free project. Pro is $25 per organization and includes unlimited projects.

Build a small app Full pricing

Hobby

One project, 1 GB RAM, 3 GB volume, 100k requests a month, SQLite, magic-link auth, and idle sleep. No card.

Pro · organization / month

$25

Unlimited projects, 5M requests, multi-region, always warm, custom domains, SSO, audit log, and snapshots. Larger resources bill by use.

Keep the source, runtime, and data.

Run it elsewhere

The Pylon runtime is open source and ships as one binary for your own server or container platform.

Use an ordinary database

Data lives in SQLite or Postgres, without a proprietary storage layer between you and your records.

Keep every source file

Export the project or connect GitHub from the start. What gets built here is not trapped here.

Build the small app your team needs.

Start with one free project and deploy it from the same place you built it.

Build a small app Read the Cloud docs