FOR AGENCIES · DRUPAL, WORDPRESS, SHOPIFY, LARAVEL, AI

White-label development your client never sees.

Subcontracting is not really a resourcing decision. It is a trust decision, and you are the one who has to defend it if the build goes sideways. The engineer you hire is writing code your client believes came from your team, and their judgement becomes your reputation the moment it ships.

Zelvatech works as the engineering arm of agencies that would rather not carry a full-time Drupal or Shopify specialist. We work inside your repository, on your board, to your definition of done. Your client keeps dealing with you, and there is nothing in the code, the commits or the deployment logs that says otherwise.

WHO HOLDS WHAT
YOU KEEP
  • The client relationship
  • The brand on the work
  • The commercial terms
  • The credit and the case study
  • Ownership of all delivered code
WE HANDLE
  • Architecture and build
  • Code review and QA
  • Migrations and upgrades
  • Deployment and handover
  • Documentation for your team
Why agencies call

Three situations, and they need different things.

Most enquiries are one of these. Knowing which one you are in changes how we would staff it and what it should cost.

SITUATION 01

The roster is full

A project closed faster than expected, or one already running has swallowed more of the team than planned. The skills exist in-house, the hours do not.

  • Usually needs someone productive inside a week
  • Overflow capacity suits this better than a fixed-scope contract
  • Works cleanly when your team writes the tickets and we clear them
SITUATION 02

Nobody in-house does this

A client wants Drupal 11, or a Shopify B2B catalogue, or an LLM search feature, and it is not something you would hire for on the strength of one project.

  • Needs depth more than hours, so a smaller team goes further
  • Usually fixed-scope with a written architecture up front
  • Often turns into the platform you then support long term
SITUATION 03

Something has gone wrong

A developer left mid-build, an offshore team delivered something unmaintainable, or a migration stalled and the client has stopped being patient.

  • Starts with an audit before anyone promises a date
  • Sometimes the honest answer is to rebuild rather than repair
  • We will tell you which one it is before you quote the client
How it actually works

We adopt your process. You should not have to adopt ours.

A subcontractor who insists on their own tooling creates a second process for your team to manage. We work the way you already work.

Inside your systems

Accounts on your Git organisation, your project board, your staging environments. We follow your branching model, your commit conventions and your review checklist.

  • Jira, Asana, Linear, Trello, ClickUp or whatever you run
  • GitHub, GitLab or Bitbucket under your organisation
  • Your coding standards and pull request template
  • Your CI pipeline, or we will build one if there is not one yet

Communication that fits

Some agencies want us in Slack and on the daily call. Others want silence and a written summary on Friday. Both work, and we agree which before the first sprint.

  • Embedded: your channels, your standups, treated as staff
  • Arms length: written updates on an agreed cadence
  • A single named engineer as your point of contact either way
  • Escalation path agreed up front, not improvised at the deadline

Nobody talks to your client

Every question and update routes through you. If a call genuinely needs technical depth, we can join it under your brand, introduced as your team, but only when you ask for it.

  • Non-solicitation written into the agreement
  • No direct contact with your client at any point
  • No approach to them during or after the engagement
  • Nothing published about the work without your written permission

Confidentiality before code

Your NDA, your master services agreement and your white-label terms get signed before we are given repository access. If you do not have templates, we can supply a starting point.

  • Copyright in delivered code assigns to you on payment
  • No licence retained and no reuse of client-specific work
  • Any generic internal tooling we bring is declared up front
  • Access revoked on the date you choose at handover
Engagement models

Five ways to buy the same engineering.

The right one depends on how well defined the work is and how long you need it for. We will say which we think fits before quoting.

MODEL 01

Dedicated capacity

A set number of engineer days reserved for you each month, invoiced monthly on a fixed date. You direct the work through your own board and it can move between your clients as your priorities change. Most agencies start at eight days a month and scale from there.

MonthlyYou set prioritiesPredictable cost
MODEL 02

Fixed-scope project

A defined build with a written scope, a fixed price and an agreed date. Best when the requirement is clear enough to specify, which after a proper audit it usually is.

Fixed priceWritten scopeAgreed date
MODEL 03

Overflow capacity

Short bursts when a project lands and the team is already committed. Nothing reserved, no monthly minimum, and we tell you honestly whether we can start in the window you need.

Short burstNo minimumFast start
MODEL 04

Rescue and recovery

Taking over a build that has stalled, overrun, or lost the developer who understood it. Always begins with an audit, because a date promised before reading the code is a guess.

Audit firstHonest scopeNo blame
MODEL 05

Time and materials

Billed by the day against an agreed rate, for work too loosely defined to fix a price to. We use it for discovery, for small support tickets under a retainer, and for the first phase of a rescue project before anyone knows what is really in the codebase. Once the scope is clear we will usually recommend moving to a fixed price, because a number your client can rely on is worth more to you than an open meter.

Day rateDiscoveryRescue phase one
Getting started

From first email to merged pull request, usually inside two weeks.

DAY 1

You send the problem, we reply

A repository, a brief, or a few sentences about what has gone wrong. The reply comes from the engineer who would run the work, inside one working day, with the questions we would need answered to scope it.

DAYS 2–4

Technical assessment, at no cost

We read the codebase and write up what is actually there: the real scope, where the risk sits, what we would build differently and why, and a price against it. Yours to keep whether or not you go ahead, and several agencies have used ours to check a quote from somebody else.

DAYS 4–6

Paperwork and access

NDA, master services agreement and white-label terms signed. Then accounts on your Git organisation and board, a walkthrough of your standards and release process, and local environments running against your codebase.

WEEK 2

First sprint, first merge

Sprint planning against your backlog, then work shipped in small reviewable branches. We deliberately pick something real but low risk for the first pull request, so your team can judge the code before anything important depends on it.

Commercial terms

The things you would otherwise have to ask for

None of this is unusual, and you would negotiate all of it anyway. Stating it up front saves a round of email and tells you how we work before you have spent anything.

Start with a paid pilot

One sprint, one module, or the first phase of a migration, scoped and priced on its own. You judge the code and the communication before committing to anything larger. We would rather lose a large project than start one that should have been tested first.

IP assigns to you on payment

Written into the agreement rather than promised on a website. We retain no licence over delivered code and do not reuse client-specific work. Any generic internal tooling we bring is declared before we start.

Your repository from day one

The code lives in your Git organisation, not ours. You can revoke our access at any point and still have everything, which is the only arrangement worth agreeing to.

Milestone-based payment

Payment tied to delivered milestones rather than elapsed time. Cross-border contract enforcement is difficult in both directions, so the practical protection for both of us is that neither is far ahead of the other.

Escalation agreed before it is needed

Who is contacted, through which channel, and what we are authorised to do without waiting for approval. Agreed at onboarding rather than improvised at two in the morning.

References on request

Including from agencies who have used us white-label. Ask them about a project that went badly, not one that went well.

Questions

What agencies check before signing.

Will you ever contact our client directly?
No. Every question, update and escalation goes through you. If a client-facing call needs technical depth, we can join it under your brand and introduced as your team, but only if you ask. Non-solicitation is written into the agreement before we see any code.
Whose name appears on the commits and in the repository?
Whatever you decide. Most agencies give us accounts on their own organisation using addresses at their domain, so the history looks like internal work. Others prefer our own named accounts for auditability. Both are normal and neither changes who owns the code.
Who owns the code we pay for?
You do, outright, on payment. Copyright assigns to the agency, not to us. We keep no licence over it and we do not reuse client-specific code elsewhere. Generic internal tooling we bring with us stays ours and is documented up front so there are no surprises later.
Can your engineers join our Slack and standups?
Yes. Some agencies embed us in their channels and daily calls and treat us as staff. Others prefer we stay out of client-visible spaces and send written updates instead. We work either way and agree it before the first sprint.
Can we resell your work at our own rates?
That is the point of the arrangement. We quote you a rate, you price the client however you price the client. We never ask what you charge and we do not publish rate cards that would undercut you.
What if we need to scale up mid-project?
Adding a second engineer usually takes a week, because they onboard into a codebase our team already knows instead of starting cold. We will tell you honestly if the work does not split cleanly, since two people on a task that suits one makes it slower rather than faster.
Do you sign our NDA or use your own?
Yours. We sign the agency's NDA, master services agreement and white-label terms as supplied. If you do not have a template we can provide one to start from, but the agency's paperwork takes precedence.
Can you work to our QA process and definition of done?
Yes, and we would sooner use yours than impose ours. Send the coding standards, branching rules, review checklist and release process at onboarding and we will follow them. If something in there will cause problems later we will say so once, and then work to it anyway.
What happens when the project ends?
You get a handover document written for a developer who has never spoken to us, the CI/CD pipeline with credentials, and a walkthrough call if you want one. Access to your systems is revoked on the date you choose. Many agencies keep a small monthly retainer afterwards, but nothing depends on it.
How do you bill white-label work?
Either a monthly rate for a reserved number of engineer days, or a fixed price against an agreed written scope. Both are invoiced to the agency, never to the end client. Before anything is signed you get a free written technical assessment covering the real scope, the risks and a price.
How big is the team?
A core team of eight engineers, scaling to around fifteen when a project needs it. Drupal, WordPress, Shopify, Laravel and front end are all covered in-house. We would rather give you the number than describe ourselves as a boutique.
When was Zelvatech founded?
2025. The company is young and the career behind it is not, which is why the portfolio is attributed to our founder rather than to the company. We think that is the honest way round, and it is why we offer a paid pilot before anything large.
Who writes the code if the founder is unavailable for two weeks?
Senior engineers on the team write most of the code already. Our founder sets the architecture, owns the review standard and is the escalation point, supported by automated checks and an AI-assisted pass that runs over every branch before a human opens it. When he is away, which is planned rather than sudden, another senior engineer takes the review role and the same standards and tooling apply. Delivery does not depend on one person being at a desk.
Can we start with something small before committing?
Yes, and we would prefer it. A single sprint, one module, or the first phase of a migration, scoped and priced on its own. You get to judge the code and the communication before either of us is committed to anything significant, and we find out whether the codebase is what everyone assumed.
Can you provide references?
Yes, on request, including from agencies who have used us white-label. Ask them about a project that went wrong rather than one that went well. That is the conversation that tells you something.
How is white-label work billed?
Either a fixed price against a written scope, or a set number of engineer days reserved for you each month and invoiced monthly. Work too loosely defined to price, such as discovery or the first phase of a rescue, is billed by the day until the scope is clear. Everything is invoiced to you, never to your client, and we never ask what you charge them.
Get in touch

Send us the one you would rather not staff.

A repository, a brief, or three sentences about what has gone wrong. Within one working day you get a reply from the engineer who would run it.

Before anything is signed, you get a written technical assessment: the real scope, where the risk sits, what we would build differently and why, and a fixed price against it. It costs nothing and it is yours to keep either way.

Reply within one working day, from an engineer Free written technical assessment NDA signed before we see code