Prerequisites
The accounts, domain plan, tools and knowledge you need before bootstrapping Lynkoria's OpenTofu deployment.
OpenTofu creates the infrastructure, but it cannot sign up for accounts, pay bills or choose your domain. Have these ready before step 1. Creating the tokens and keys is part of step 2.
Accounts
| Account | Why | Notes |
|---|---|---|
| DigitalOcean (with billing) | App Platform, Droplets, managed databases, Spaces, container registry | Request a higher Droplet limit if your account cannot create dedicated-CPU (c-*) sizes |
| DigitalOcean Container Registry | Holds the API and scanner images | Reuse an existing registry; its plan must allow two more repositories (Basic or higher) |
| Cloudflare | DNS for the domain, proxy and TLS for the API, Access, Pages | Free plan works; the domain's zone must be on Cloudflare |
| Cloudflare Zero Trust (recommended) | Access, with MFA, in front of the console and /v1/admin; optional (enable_access) | Free for small teams; needs an identity provider (Google, GitHub, Okta, …) or one-time PIN |
| GitHub | Runs CI and deployments; holds the deployment secrets | Admin access to the repository settings |
| Firebase / Apple Developer | Optional: push notifications per product | Only when a product enables push |
| Monitoring provider | Optional: Prometheus remote write, alerting | Grafana Cloud, a self-hosted Prometheus/Mimir, or any remote-write endpoint |
Lynkoria sends no email; console accounts are created by operators with a CLI, so no SMTP provider is needed.
Domain plan
The hostnames are set in infra/opentofu/platform/environments/*.tfvars. The repository uses:
| Purpose | Production | Staging |
|---|---|---|
| API | api.lynkoria.com | api-staging.lynkoria.com |
| Admin console | console.lynkoria.com | console-staging.lynkoria.com |
| Docs / landing | lynkoria.com | staging.lynkoria.com |
| LiveKit | rtc.lynkoria.com | rtc-staging.lynkoria.com |
| TURN | turn.lynkoria.com | turn-staging.lynkoria.com |
| Scanner (private) | scanner.internal.lynkoria.com | scanner-staging.internal.lynkoria.com |
Also list every product web origin (for example https://app.yourproduct.com) that will call the API from a browser. They go into allowed_origins and the bucket's CORS rules. Wildcards are refused.
Tools
On the machine you bootstrap from (Ubuntu shown; on macOS use Homebrew equivalents):
| Tool | Version / note | Check |
|---|---|---|
| Git | any recent | git --version |
| Node.js + pnpm | Node from .nvmrc (26), pnpm 12.18.0 | node --version |
| OpenTofu | the version in infra/opentofu/.opentofu-version | tofu version |
doctl | DigitalOcean CLI | doctl version |
| AWS CLI v2 | S3-compatible client, to create the state bucket | aws --version |
jq, psql, curl, openssl | used by scripts and checks | jq --version |
| Docker | optional: local builds and the local stack | docker version |
Install OpenTofu from opentofu.org, doctl from DigitalOcean and the AWS CLI from AWS.
sudo apt-get update
sudo apt-get install -y git jq postgresql-client openssl curlGitHub Actions installs its own OpenTofu, doctl and Wrangler; you only need these tools locally for the bootstrap, troubleshooting and the occasional manual operation.
What you do not need
You do not run the LiveKit generator, create Droplets, databases, buckets, DNS records, firewall rules or Access applications by hand, and you do not generate application secrets. OpenTofu does all of that. The full split is in status and gaps.