DAEDALUS

Custom software, built fast.

About Us

We think companies should own the software that runs their business. That used to be a luxury only large enterprises could afford. AI has become a force multiplier for software engineering, allowing us to bring custom software to small and medium sized businesses.

Building your own software gives you complete control and flexibility of every digital aspect of your business. We should know, because we have created or borrowed from open source almost all the software our business runs on, and we run it all on computers that we own. Instead of submitting a support ticket to software you rent, you can implement a feature next week in software you own. Stop building your business around SaaS, build software around your business.

We have built our business to be AI-native from the ground up (custom software, automated workflows, company MCP with tools and agentic cross-communication) and constantly evolve with the cutting edge. We are on a mission to help your company get there too.

Our Constitution

Let’s Talk

Have a project in mind?
We’d love to hear about it.

BOOK A CALL
OpenAI Select Partner

Team

Hunter Warburton

CTO

Hunter Warburton

Software engineer with a background in infrastructure at scale, automation systems, and local AI deployments. Physics degree from Colgate University. Passionate about lean, self-hosted solutions. Recently gave his personal LLM tool access to his whole genome.

Johnny Gabriele

CEO

Johnny Gabriele

Serial entrepreneur and digital transformation specialist. Co-founder of the award-winning political strategy firm, HZQ Consulting. NYU-trained organizational psychologist. Ex brand strategist for Google, Netflix, NBC, Electrolux, Pfizer. Loves to help founders navigate complexity, and cats.

Joe Koenig

COO

Joe Koenig

Business consultant and AI implementation partner with a background in insurance and economic analysis. Former Bankless Consulting tokenomics analyst and founder of Risk Meridian. Joe helps clients assess business opportunities, map workflows, and turn research into working software. He works with teams to define requirements, review results, and put new tools into use.

Elliott Conway

Software Engineer

Elliott Conway

Builds practical tools and system integrations. Elliott works directly with clients to understand their workflows, build working prototypes, and refine them through feedback. Based in Colorado.

Ethan Lockshin

Advisor

Ethan Lockshin

Business development at NVIDIA, guiding AI startups through the Inception program. Previously sold AI inference at Groq until its 2025 exit to NVIDIA, led sales at the early AI agent platform Kubiya.ai, and was a top-ranked account executive at CircleCI. Based in Denver.

FAQ

How much does custom AI software cost?

Every Daedalus engagement starts with a package of 20 credits for $10,000, or $500 per credit. About five credits cover the blueprint, in which we interview your staff, map how the work gets done and produce a credit menu that ranks the improvements worth making and puts a credit range on each. For example, a weekly report that compiles itself from your existing systems might be priced at 4 to 6 credits, or $2,000 to $3,000. You choose what to build with the remaining credits, and larger systems take additional packages. If you stop after the menu, you keep the blueprint and we refund the unused credits.

How does an engagement with Daedalus start?

Every Daedalus engagement starts with a blueprint, which is part of the first credit package. We interview your staff and map how the work gets done, because the handoffs and exceptions that cost the most rarely come up when someone describes the feature they want built. The credit menu that follows includes only improvements that save money, free up staff time or improve the customer experience, and each one is priced before you commit. You choose what we build, we test it with your team, and you receive the code. If you are unsure whether a blueprint is worth doing, send us the workflow that wastes the most of your team’s time and we will tell you.

Why does Daedalus charge in credits instead of by the hour?

Daedalus prices work in credits because AI coding tools have made hours a poor measure of what you receive. A task that used to take four hours can now take fifteen minutes, and neither number describes what the result is worth to your business. A credit prices the piece of work itself, each item on your credit menu carries a range set before you commit, and we refund credits you do not use.

How long does a custom software project take?

A Daedalus engagement starts with the blueprint, and the timeline after that depends on what you choose from the credit menu. A focused automation is often built within a month. Custom software that replaces or connects core systems has taken three to five months, because access, vendor approvals, testing with your staff and training take calendar time that faster coding does not shorten. You see working pieces throughout, so problems surface while they are still cheap to fix.

Should we build custom software or buy an existing tool?

You should buy an existing product when it already handles your process well, and Daedalus will tell you so. Our constitution commits us to comparing off-the-shelf, custom, rule-based and AI options before recommending one, and to recommending another provider when that is the better answer. Custom work earns its cost where your process differs from what generic tools assume. Software you own also becomes an asset of your business, which a subscription you rent never does.

Who owns the code and data when the work ends?

You own the code and data at the end of a Daedalus engagement, and we return or destroy anything we hold and confirm that in writing. We design for portability, so another capable team can understand, run and change the system without asking us for permission. Article VII of our constitution sets out this commitment.

Will our data be used to train AI models?

Daedalus does not use client data to train or improve AI models, and client content is not opted into training with the AI providers we use. For sensitive documents, the AI runs on open-weight models hosted on GPUs Daedalus owns, so those documents never pass through a cloud AI provider. Where you prefer your own cloud, we can run models through a service such as AWS Bedrock inside your account, under your agreements with that provider.

How does Daedalus handle sensitive or regulated data?

Daedalus runs each client’s application and database separately, either on hardware we own behind a single hardened server that is the only part reachable from the internet, or entirely inside the client’s own cloud account. For healthcare clients we sign business associate agreements, and our document pipeline processes protected health information in memory and writes it back to the client’s own system instead of storing it. We maintain a written annual security risk assessment, based on NIST SP 800-30 and mapped to the HIPAA Security Rule, with a tracked remediation plan kept under version control.

Blog

Worn book cover showing a sealed mechanism opening above small workshops, titled If It's Useful, It Won't Stay Closed.

If It’s Useful, It Won’t Stay Closed

By Hunter Warburton

& muse-wonder-full-1.0

A worn vermilion cloth field manual with a searchlight woodcut and cream title lettering.

You’re not too small to get hacked

By Hunter Warburton

& GPT-6 Astra

A worn mint book jacket with the article title above an archive cabinet and an open coral side door.

Is Your AI Provider Stealing Your Data?

By Johnny Gabriele

& Fable 5 and GPT 5.5

Antique technical book cover with an empty automation suit titled Superpowers, Made to Measure

Superpowers, Made to Measure

By Joe Koenig

& Fable 5

Antique book cover with an expanding intelligence diagram titled The Token Panic

The Token Panic

By Hunter Warburton

& Opus 4.6

Grid of silhouetted figures with unique patterns on a worn book cover

Barriers Eroded

By Hunter Warburton

& Opus 4.6

Moonlit labyrinth at sea on a dark magazine cover

The Bet

By Hunter Warburton

& GPT 5.5