IT services · Technology consulting
MyraClean Gebäudereinigung GmbH plans, builds and supports business software: custom applications, web platforms, cloud infrastructure, integrations between systems that were never designed to cooperate, and the reporting layer that turns daily operations into something you can read.
Who we are
MyraClean Gebäudereinigung GmbH operates as a technology practice for organisations whose software has become central to how work actually gets done. That includes internal tools nobody planned for, spreadsheets that quietly became systems of record, and platforms that grew faster than the documentation around them.
The approach is deliberately unglamorous. Before proposing a build, we establish what already exists, who depends on it, and what would break if it changed. Scope is written down. Assumptions are labelled as assumptions. Where the honest answer is that a smaller change would solve the problem, that is the recommendation.
Work is carried out in short, reviewable increments so that progress is visible in a running system rather than in status reports. Nothing on this website describes past client engagements; scenarios shown further down are illustrative examples of how work is structured.
Capabilities
Capabilities are described in terms of the problems they address. Most engagements combine two or three of them, because a new application usually needs somewhere to run, something to integrate with, and someone to keep it healthy afterwards.
Applications shaped around a specific operational process rather than around a generic product roadmap.
Browser-based systems with defensible performance, accessibility and interface behaviour.
Environments described in code, with deployment, observability and cost visibility built in.
Reliable movement of data between applications, including error handling and replay.
Removal of repetitive manual steps, with audit trails and human checkpoints where they matter.
Models, pipelines and reporting surfaces that make operational reality legible.
Dependency updates, monitoring review, small changes and a defined path for incidents.
01 — Custom software
Custom development is appropriate when a process is specific enough that off-the-shelf software would need to be bent out of shape to fit it. Typical artefacts include internal operations tools, scheduling and dispatch systems, inspection and quality records, quoting logic, and back-office applications that replace an accumulation of spreadsheets.
02 — Web engineering
Web applications are judged by what happens on an average device, on an average connection, in the middle of a working day. That means measured page weight, predictable layout, keyboard operability, sensible error states and forms that do not lose input.
Server rendering or prerendering where content should be visible immediately; images sized and lazily loaded below the fold; fonts loaded without blocking the first paint.
Semantic structure, meaningful heading order, descriptive alternative text, visible focus states, sufficient contrast, and respect for reduced-motion preferences.
Clear separation between server data, cached data and interface state, so that refreshing, retrying or navigating never leaves the screen in an ambiguous condition.
Layouts tested at narrow, tablet and wide widths, with long content, missing content and unusually long labels all considered.
03 — Cloud
Modernisation work usually begins with an inventory: what runs where, who can reach it, how it is deployed, and what would happen if the machine it depends on disappeared. From that inventory we propose a target state that your team can operate — not the most fashionable architecture available.
Declarative definitions for networking, compute, databases and access, so staging and production do not drift apart.
Automated pipelines with build artefacts, migrations and a documented way back to the previous release.
Structured logs, metrics and alerts chosen so that an alert means something actionable.
Resources tagged and reviewed, so infrastructure spend can be explained line by line.
04 & 05 — Integration and automation
Integration work is mostly about failure. Two systems exchanging records on a good day is straightforward; the engineering is in duplicate messages, partial writes, schema changes, expired credentials and outages on the other side. We design for those cases explicitly, with idempotent operations, retry policies, dead-letter handling and a way to replay what was missed.
Automation follows the same logic. A manual process is first documented as it is actually performed, including the exceptions people handle informally. Only then is it automated, with human approval retained at the points where judgement is required, and with an audit trail that records what ran, when, and on which record.
Interfaces
REST and webhook endpoints, message queues, scheduled file exchange, database replication.
Guardrails
Schema validation at boundaries, contract tests, alerting on backlog growth.
Traceability
Correlation identifiers so a single record can be followed across systems.
06 — Security-conscious practice
No engineering practice can promise that a system is secure, and this page makes no such claim. What can be described are the habits applied by default.
Least-privilege roles, separate credentials per environment, no shared accounts, access reviewed when it is granted rather than long afterwards.
Configuration and keys held in a secret store, never committed to source control, rotated when people or suppliers change.
Locked versions, automated vulnerability alerts, and a scheduled window for applying updates rather than an annual scramble.
Encrypted transport, minimal collection, retention limits agreed in writing, and logs designed not to accumulate personal data by accident.
07 — Data and analytics
Most reporting problems are definition problems. Before building dashboards, we agree what each measure means, which system is authoritative for it, and how edge cases are counted. That agreement is written into the model rather than kept in someone’s head.
Delivery
Stage 01
Interviews with the people doing the work, review of existing systems and data, constraints and deadlines recorded. Output: a written problem statement and options.
Stage 02
Scope divided into increments, technical approach chosen, risks named with mitigations. Output: a plan with a first releasable slice identified.
Stage 03
Short iterations with working software at the end of each. Code review, automated tests and a running staging environment throughout.
Stage 04
Functional checks against agreed criteria, performance review on realistic data volumes, accessibility pass, and a security review of access and secrets.
Stage 05
Deployment through the automated pipeline, documentation of environments and runbooks, and a walkthrough with the people who will operate the system.
Stage 06
Agreed cadence for dependency updates, monitoring review and small changes, with a defined route for reporting problems.
Quality
Test effort is spent where failure would be expensive, not spread evenly for the sake of a coverage figure. Business rules, calculations, permissions and integration boundaries are tested closely; presentational details are checked by review.
Unit tests for logic, integration tests against real database behaviour, and a small set of end-to-end tests covering the paths a user cannot afford to lose. The pipeline blocks a release when they fail.
Every change is read by another engineer, with attention to naming, error handling, data migrations and whether the change is reversible.
Behaviour is checked against representative volumes rather than a handful of sample rows, so slow queries appear before users find them.
Reported problems are reproduced, covered by a test, then fixed — so the same failure does not return quietly in a later release.
Illustrative examples — not client work
The following scenarios are constructed examples written for this website. They do not describe completed projects, real organisations, or measured results.
Example A
A hypothetical operations team coordinates daily assignments across several shared spreadsheets. Discovery would map the current process and its informal exceptions. A first increment might deliver a single assignment view with a proper data model and an audit trail, followed by conflict detection and printable day sheets. The spreadsheets would remain readable during the transition rather than being switched off in one step.
Example B
A hypothetical application runs on one ageing server with manual deployments. The work would begin by describing the environment in code and building a repeatable deployment, then adding monitoring, then extracting the most fragile component. Behaviour would be captured in tests before anything is moved, so the migration can be verified rather than hoped for.
Example C
A hypothetical finance and operations group receives conflicting numbers from three tools. The engagement would define each measure and its authoritative source, build scheduled pipelines into a single model with freshness tests, and publish a small set of dashboards plus spreadsheet exports. Definitions would be documented alongside the model.
Collaboration
Most project difficulty is communication difficulty. These principles apply to every engagement regardless of size.
Scope, decisions and open questions live in a shared document rather than in scattered messages.
A brief written summary of what changed, what is next and what is blocked, at an agreed cadence.
Progress is shown in a running staging environment so feedback is grounded in something real.
Where speed, cost and durability conflict, the options and consequences are written down before a decision.
Whoever will operate the system participates during the build, not only at handover.
Slippage, misunderstandings and technical dead ends are raised as soon as they are recognised.
Questions
Contact
MyraClean Gebäudereinigung GmbH works with organisations that would rather understand their software than merely own it. If that describes the situation you are in, a short written description of the problem is enough to begin a conversation.
MyraClean Gebäudereinigung GmbH
audreycurt77@gmail.com
myraclean.com
Please include the outcome you need, the systems already in use, any deadline, and the person who can answer technical questions. Contact details are shown as plain text.