01The problem it solves
Node.js is a runtime, not a framework. Choosing it does not settle the architecture: you still have to decide how the service is structured, how inputs are validated, how retries work and how what happens is observed.
Its advantage is concurrent input/output, not heavy computation. If the work is CPU-bound, Node competes badly and another runtime is a better fit; if the problem is only connecting tools, it may be automation rather than a service of its own.
Adding a new runtime also adds operation: versions, dependencies, the process, logs and limits. That cost is accepted from the start or it turns into debt.
If the problem is solved with a direct call or an automation tool, standing up a Node backend only adds a system to maintain. The service is justified by what it does, not by the runtime it uses.
02When it adds value and when it doesn't
Can add value
- Several third-party services have to be integrated through APIs, webhooks and different formats.
- The product needs real-time work: notifications, shared states or events.
- The service is small, decoupled and has a clear responsibility.
- The team works in JS/TS and can maintain the runtime and its dependencies.
- The load is input/output: many connections or network waits, little CPU.
- Laravel or WordPress do not cover the API and adding Node avoids forcing the other stack.
Doesn't add value
- It is the default corporate website: WordPress handles it better with less operation.
- The task is CPU-intensive, where Node is not the best choice.
- The team only works in PHP and does not want to take on a new runtime.
- Laravel already covers the API at no extra cost: a second backend is not added.
- The problem is orchestrating a process across tools: that is automation, not a service of its own.
- No one can operate the process, logs, limits and dependencies in production.
The four decisions
-
01
Keep WordPress
If the case is content, standard ecommerce or a simple integration a plugin handles, WordPress is enough. Node is not installed inside the site and does not replace it.
-
02
Solve it with Laravel
If the centre is a PHP application with a panel, ORM, queues and its own rules, Laravel covers the API without adding a second backend. Node is considered when integration or I/O concurrency is the core, not as an alternative by preference.
-
03
A dedicated Node.js service
If the service is about integration, webhooks or real time, and the team works in JS/TS, Node fits. It means taking on a new process and deployment, with its own limits and monitoring.
-
04
Do not add another runtime
If the PHP API already covers the case, the team cannot operate Node or the problem is CPU-bound, it is not added. Two backends for the same domain duplicate deployment, security and maintenance.
03What we do
Node.js isn't the service: the service is what gets built with it. These are the usual pieces, each one with its scope and its cost.
-
01
API with a contract and input validation.
Go to Custom applications -
02
Webhooks with verification, retries and idempotency.
Go to Custom applications -
03
Queues for work that must not block the response.
Go to Custom applications -
04
Integrations with the third-party services involved.
Go to Custom applications -
05
Authentication and authorisation for the case.
Go to Custom applications -
06
Tests for critical flows and foreseeable failures.
Go to Custom applications -
07
Logs, metrics and limits to know what happens in production.
Go to Custom applications -
08
Deployment, documentation and handover with the agreed repository.
Go to Custom applications
04Performance and maintenance
Can be measured
- Latency of responses and of calls to external services.
- Real concurrency the process sustains before it degrades.
- Configured limits and timeouts, and what happens when they are reached.
- Errors, retries and logs from integrations in production.
Can't be promised
- That real time works without measuring the connection, the load and behaviour on slow networks.
- That Node is always faster than another backend: it depends on the type of load and the code.
- A number of connections or events per second without having tested it under equivalent load.
Maintenance
Update Node and the dependencies, review the lockfile and the npm chain, and maintain the process, logs and monitoring. Node is not a framework: every piece chosen is maintained. The cost is agreed with the project.
Price and timeline
Quoted after diagnosis. It depends on integrations, volume, authentication, tests and operation. No rate is published while the scope is not defined.
See the speed and WPO service05Proof: stated capability
Declared capability. There are no publishable cases or verified metrics for Node.js services on this project, and evidence per technology is still pending.
What I can explain is the reasoning: how the service is bounded, how integration failures are tested and what is measured before talking about capacity.
This page describes when Node fits and when it does not; it does not claim specific experience or promise performance that has not been measured.
See the projects06Frequently asked questions
Node.js or Laravel?
It depends on the team and the problem. Laravel fits when the centre is a PHP application with a panel, ORM, queues and its own rules. Node fits when the centre is integration, real time or input/output concurrency, and the team works in JS/TS. Both are not chosen for the same problem without need.
When does Node.js make sense?
When services have to be integrated, webhooks handled, real-time work kept running, or a small decoupled API served with a lot of network wait. Not when the website is content, the load is CPU or the PHP backend already solves it.
What maintenance does it need?
Node and dependency versions, a lockfile, package audit, the process, logs and monitoring. Node is not a framework: the pieces have to be chosen and maintained. The cost is agreed with the project.
Is it secure?
Security depends on the application: validation, authorisation, secret management, limits and dependencies. No plugin covers it. The npm chain and permissions are reviewed before deployment.
How much does it cost?
Quoted after diagnosis. It depends on integrations, volume, authentication, tests and operation. No rate is published while the scope is not defined.
How does it connect to WordPress?
As a separate service that exchanges data through an API, or as an integration layer for WordPress and n8n. It is not installed inside the corporate website and does not replace it.