Skip to content

Ship the app.
Call a model.
One platform.

Connect a Git repo or a container image. Add a database. Then point any OpenAI-compatible client at the same endpoint, with the same key.

Start building Explore the platform

Deploy

Repository, Dockerfile, or image

Data

Five engines, private to your project

Intelligence

OpenAI-compatible chat, embeddings, images, video

Operate

Logs, state, and deploy history

Your stack, already known

Bring your own code

Two ways in, one way out

A repository, or an image you already built. Maxlayer meets your code where it lives.

Connect GitHub or GitLab

Authorise once, pick a repository and a branch. We set up the deploy key and webhook; when it is active, pushes create deployments.

Container image

Already built it? Point us at the image, skip the build, and run the container you made.

4 steps, 0 servers

Empty account to live URL

Four steps, and none of them involve provisioning anything — or opening an account somewhere else for the fourth.

  1. 01

    Point at your code

    Authorise a repository or name an image. Automatic detection reads the project, or you can choose a Dockerfile or static output.

  2. 02

    Feed it the environment

    Add variables one at a time or paste a whole .env. Flag the sensitive ones secret and they stop reading back.

  3. 03

    Put it in service

    Start the container on the port you choose. Add a hostname, verify its DNS, and request a certificate without leaving the dashboard.

  4. 04

    Call a model with the same key

    Point an OpenAI-compatible client at the inference endpoint. The organization key that deploys your app authorizes its chat and embedding calls too.

What runs here

Run web apps , APIs , AI features , long-running services , static sites and internal tools — each one a repo, a build, and a container.

An app is one deployable thing: a site, an API, a worker, the service behind a chat box. Most teams run several. They all work the same way, which is the point. The ones that call a model use the same endpoint and the same key as the rest.

Web apps and SaaS

Next.js, Rails, Django, Laravel. Build from source, route to the port your process listens on, and add a domain when you are ready.

APIs and backends

Ship a REST or GraphQL service, then hand CI a key scoped to exactly that one job.

AI features and agents

A retrieval service, a chat endpoint, a summariser running as a worker. The code runs here, and so does the endpoint it calls.

Long-running services

Queue consumers and workers run as their own container, with their own build settings, plan, and deploy history.

Static sites and docs

Static output skips the runtime entirely. Built from source, served straight, nothing left running.

Internal tools

Operations panels and back-office services, with platform access scoped per person by membership role.

Somewhere to put it

Data that lives next to your app

Add one in the dashboard and it comes up beside your apps, on your project's own network. Attach it and the connection string is written in as a secret — no copy-paste, no credentials in a chat thread.

PostgreSQL

The default answer, and usually the right one.

MySQL

Widely supported, widely understood.

MariaDB

MySQL-compatible, community maintained.

MongoDB

Document store. Schema on write.

Redis

Cache, queues, sessions. Not a store of record.

Databases start private to your own applications — logs, scheduled backups, and a port you can open deliberately if you need to reach one from your laptop. They bill by the hour like everything else, and stop costing anything the moment you stop them. How databases work here , or see the tiers — including what we do not do yet.

Intelligence

Everywhere else, the AI half is a second vendor

Here it is part of the platform you already pay for — one endpoint beside your apps, and a public catalogue that publishes every model's rate before you send a request.

No second SDK

Chat completions, embeddings, images and video on the OpenAI request and response shapes, through the clients your project already uses.

No second key

The organization key that deploys your apps and drives your automation authorizes inference. Cursor, Continue, Cline, Aider, and Zed take the same one.

No second bill

Tokens draw from the prepaid credit that pays for your apps and databases, and the cost shows up beside theirs.

openai.ts 
const client = new OpenAI({  baseURL: "https://inference.maxlayer.cloud/v1",  apiKey: process.env.MAXLAYER_API_KEY,});same client, new endpoint
Explore AI inference See the models and their rates Read the compatibility guide

The boring 90%

Everything a running system needs

No proxy to configure, no certificates to chase, no runbook to write. Handled before you go looking for it.

Build on push

Connect a branch and let a working webhook start deployments for you. Rebuild by hand whenever you need one.

Custom domains and TLS

Verify DNS, point a hostname at your app, and request a Let's Encrypt certificate from the same place.

Secrets that stay secret

Mark a variable secret and it becomes write-only to the dashboard and API after you save it.

Logs with context

Follow build output, inspect runtime logs, and read back the record for every deployment.

Deploy history

Every deployment on the record — status, duration, and who asked for it.

Dashboard or API

Create apps, update environments, trigger deploys, read token usage, and manage the lifecycle from either surface.

Not a solo sport

Built for more than one person

Everyone gets the access they actually need and nothing past it. No shared login to rotate the week somebody leaves.

  • Four roles, clearly separated

    Viewers read, developers deploy, admins run the team, owners own the account.

  • Invitations, not shared passwords

    Email an invite with the role already attached. It expires if nobody uses it.

  • API keys scoped to a team

    Hand CI a key carrying exactly the role it needs. Revoke it in one click.

ci/deploy.sh 
# deploy from CI with a scoped key$ curl -X POST \  https://api.maxlayer.cloud/v1/orgs/$ORG/apps/$APP/deploy \  -H "Authorization: Bearer $MAXLAYER_API_KEY" \  -H "Content-Type: application/json" \  -d '{"title":"Release v2.4.0"}'{ "id": "b3f8c2a1-…", "state": "deploying", "accepted": true }

Ship something today

Connect a repository and watch it build, then call a model with the same key. Apps from $8 a month, tokens by the million, nothing to cancel.

  • GitHub + GitLab
  • Container images
  • OpenAI-compatible AI
  • Pay as you run
Get started Read the docs