API

HTTP API for projects, apps, and the rest of the platform.

Every resource you see in the dashboard has an HTTP endpoint. Use this from the CLI, CI, scripts, or any client that can send JSON.

Authenticate

Send a bearer token on every request:

httpAuthorization: Bearer nodion_…

There are two kinds of token:

  • Personal API keys (nodion_…) act as you. You choose scopes and can limit the key to specific projects. Create them under Account → Security. The token is shown once.
  • Project keys (nodion_proj_…) belong to one project, not a person, so CI keeps working when someone leaves. Create them on the project’s API keys tab.

Scopes on a personal key:

  • read: list and fetch resources
  • write: create, update, and delete
  • deploy: trigger deployments
  • env:read / env:write: environment variable values
  • logs:read: application logs

Reads need read, writes need write, unless an operation asks for a tighter scope (for example deploy). Project keys are not scoped. They can do anything in that project. Billing and 2FA stay in the dashboard; API keys cannot call those.

Pick a project

Nested routes look like /v2/projects/{id}/…. If you work in one project, use the short path and send the project id as a header:

httpX-Nodion-Project: prj_…

First request

bashcurl https://api.nodion.com/v2/applications \
  -H "Authorization: Bearer $NODION_TOKEN" \
  -H "X-Nodion-Project: $NODION_PROJECT"

Trigger a deploy (needs the deploy scope on a personal key):

bashcurl -X POST https://api.nodion.com/v2/applications/$APP_ID/deployments \
  -H "Authorization: Bearer $NODION_TOKEN"

Lists, errors, and limits

Collection responses look like { "data": [ … ], "next_cursor": null }. Pass ?limit= and the previous next_cursor as ?cursor= to page.

Errors use RFC 9457 (application/problem+json) with a stable code and a request_id. Quote that id to support. The public catalog is GET /v2/errors.

Key-authenticated calls are rate limited per account. Responses include RateLimit-Limit, RateLimit-Remaining, and RateLimit-Reset. A 429 also sends Retry-After. Every response has X-Request-Id.

For request bodies and a try-it console, use the interactive docs. The pages below are the same endpoints, in this site’s look.

Endpoints

Generated from the public OpenAPI document. 298 operations, grouped as in the spec.

Identity

Validate a token, discover reachable projects, and list error codes. 4 endpoints.

Projects

Projects, the container every other resource belongs to. 7 endpoints.

Applications

Deployable applications and their settings, logs and metrics. 19 endpoints.

Previews

Ephemeral preview environments for a branch or pull request. 5 endpoints.

Environment

Application environment variables. 3 endpoints.

Deployments

Build-and-run deployments and auto-deploy. 7 endpoints.

Processes

An application's process types and their scaling. 6 endpoints.

Cron jobs

Scheduled one-shot commands on an application's active release. 10 endpoints.

Domains

Custom domains for applications. 8 endpoints.

Databases

Managed databases, replicas, endpoints and attachments. 43 endpoints.

Volumes

Persistent storage volumes. 11 endpoints.

Object storage

S3 buckets and access keys. 27 endpoints.

Registered domains

5 endpoints.

DNS

Authoritative DNS zones and records. 23 endpoints.

CDN

CDN zones, custom domains and cache purging. 16 endpoints.

Email

Email sending domains and relay zones. 40 endpoints.

Observers

Uptime and metric monitors with notifiers. 16 endpoints.

Integrations

Git provider integrations and shared repositories. 25 endpoints.

Registries

Nodion container registries and per-project robot accounts. 11 endpoints.

AI

Model catalog and per-project usage. Inference itself is the separate OpenAI-compatible API. 3 endpoints.

Reference

Read-only catalogs (regions, packages, engines, presets). 9 endpoints.

Give your agent a cloud.

Earn up to 15 € in credit when signing up.

Continue with GitHub

or with email

Stay in the loop.

Product news, no spam. Unsubscribe anytime.