Built by following
the problem.

A company for building technology across different problem spaces.

Aetras develops products, software systems, infrastructure, and experiments where a purpose-built technical approach makes sense.

The company is deliberately not defined by a single product category. Current work spans web applications, distributed systems, sound and media, realtime software, infrastructure, and exploratory technology.

Projects can begin as products, internal requirements, experiments, or individual technical problems. Useful components can then become foundations for other work.

Product development

Start with an existing solution. Build differently when it stops fitting.

The work that preceded Aetras developed around practical product requirements rather than a fixed technology roadmap. Existing platforms and tools were used where they solved the problem well enough.

When requirements became specific enough that adapting a general-purpose system introduced more limitations than it removed, the work shifted toward purpose-built software.

That transition established an approach that remains central to Aetras: understand the problem first, reuse what already works, and build the parts that genuinely need to be different.

Sound assets

A general storefront became a purpose-built sound platform.

Endalon began with existing commerce platforms. Sellfy was followed by WordPress and WooCommerce, but the requirements of a growing sound library increasingly conflicted with systems designed around conventional products.

Sound needed to be browsed, previewed, described, filtered, stored, and managed differently. Instead of continuing to extend a general-purpose storefront, Endalon moved to a custom application designed around sound assets themselves.

From application to operation

Building products created a need to operate them.

As the software grew, its requirements expanded beyond application code into deployment, databases, object storage, networking, containers, reverse proxies, continuous delivery, and production operations.

Aetras operates infrastructure with responsibilities separated across application compute, persistent data and storage, private networking, and independently provisioned inference workloads.

Research through implementation

Not every project needs to become a product.

Experimental work gives Aetras room to investigate technical ideas without requiring a product or commercial use case at the outset.

An experiment may develop into a standalone system, contribute technology to another project, establish reusable infrastructure, or simply answer the question that started it.

Experiment → System

One capability became an independent system.

Aetras Transcription developed from broader experimentation with voice and sound. Speech-to-text was isolated into a standalone capability that could develop independently and support more than one future workflow.

The resulting architecture combines browser-side media normalization, an ASP.NET API, PostgreSQL-backed job coordination, independent workers, Whisper inference, worker leases and recovery, and incremental transcription results.

A broader company

Endalon became a product.
Aetras became the company.

As the work expanded beyond a single sound platform, the original company identity no longer represented the full scope of the technology being developed.

Aetras was established in 2026 as the broader company. Endalon continues as a sound-focused product, while Aetras provides a home for software, infrastructure, experiments, services, and future products across different fields.

Products, engineering, and experiments under one name.

Aetras develops its own products and experimental technology while also taking on focused software and technical work for businesses and other projects.

The company remains intentionally small. Work is selected around problems where direct technical involvement and a focused engineering approach are useful.

C# and .NET form the core application stack across Aetras. PostgreSQL, modern browser technologies, and Linux-based infrastructure support much of the current work, with other technologies introduced where a project benefits from them.

Start with what needs to work.

A technical specification is not required before getting in touch. A problem, inefficient workflow, missing capability, or product idea is enough to begin understanding what the system actually needs.

Work can range from focused applications and internal tools to audio and media systems, automation, realtime software, distributed processing, and supporting infrastructure.

Projects are scoped around their actual requirements and responsibilities. Where a project requires expertise outside the appropriate scope of Aetras, that should be identified rather than hidden behind unnecessary complexity.