Custom software – your code, your freedom.

Under the MIT licence, nokkela.app hands over the complete, documented source code of the software it builds, and the client may use, modify, extend and pass it on with no licence fees and no permission required. MIT is a permissive open-source licence, not a transfer of copyright: the copyright notice and licence text stay with the code. You can continue development in-house or commission another provider at any time.

A nokkela.app project runs in four steps: discovery, prototype, build and handover. Discovery sharpens the requirements and assesses feasibility and effort; the prototype is a fast, clickable proof of concept working with real data, so you can judge the direction before the main build. Development is iterative, test-driven and transparently traceable. At handover you receive the source code under the MIT licence, documentation and training; operation afterwards is optional with us.

nokkela.app builds on an open, established stack: Python with Django, FastAPI or Flask plus Celery for asynchronous tasks, NodeJS and TypeScript, and React, Vue or Next.js with Tailwind CSS and Vite on the front end. Data lives in PostgreSQL, MariaDB, MongoDB or Redis, with SQLAlchemy for relational access. Interfaces follow REST, GraphQL, OpenAPI or WebSockets, secured with OAuth2/OIDC and Keycloak. We ship on Docker, Kubernetes, Nginx and Linux via Git and GitLab CI, tested with pytest and Playwright.

Modernisation at nokkela.app means replacing legacy systems, migrating data and connecting what already exists without a break in day-to-day work. In practice that combines migration and refactoring, connecting ERP, CRM and line-of-business systems over stable interfaces and automating steps that were manual before. Because the prototype runs on real data early, migration paths are tested before the switch rather than after it.

nokkela.app delivers the source code under the MIT licence together with documentation and training, so after handover your own team or another provider can operate the software. If you would rather not operate it yourself, hosting and operations are available through nokkela.host on Nokkela infrastructure in Europe, processed to GDPR standard. On-premise deployment on your own hardware is equally possible: the stack is Docker, Kubernetes, Nginx and Linux, not a proprietary platform.

Yes – nokkela.app builds AI features directly into custom software, together with nokkela.ai for the model side and nokkela.io where agent systems and process automation are involved. Typical additions are search and retrieval over your own documents, classification, extraction and assistance inside an existing workflow. Because the AI can run on infrastructure in Europe or on your own hardware, prompts and documents need not leave your environment.

The handover from nokkela.app consists of the complete, documented source code under the MIT licence, plus training for the team that will take it on. Maintainability comes from the way the code is produced: development is iterative and test-driven, with pytest and Playwright covering back end and browser, versioned in Git and built through GitLab CI. Deployment is containerised with Docker and Kubernetes, so the same application can be moved between environments.