Blacksky Labs

Building things
worth building.

An indie AI lab. Small team, big ideas. We make products at the intersection of AI, civic tech, and creative tools — thoughtful, unusual, and actually useful.

Who we are

A small lab
doing big things.

Blacksky Labs is an indie AI studio — the kind of place where curiosity leads and polish follows. We're a small, focused team building in the open, iterating fast, and shipping things that actually work.

We don't have a corporate agenda. We have a genuine interest in what AI can do when it's applied carefully — to civic accountability, creative expression, novel interfaces, and the edges where things get interesting.

7+
Products shipped
3
Active domains — AI, civic tech, creative tools
Ideas in the queue
What we do

We build at the
edges of things.

AI Systems

AI beyond the obvious. We're interested in the parts of the field that aren't crowded — swarm intelligence without neural networks, generation pipelines that think before they paint, retrieval systems that actually retrieve what matters. The common thread is treating AI as a craft, not a checklist.

Civic Tech

Technology in service of accountability. Government data is technically public and practically buried. We build tools that pull it into the open — voting records, ideology scores, money trails — and put them where ordinary people can use them, free.

Creative Tools

Tools for making, by people who like making things. We build for artists, builders, and anyone bringing something into the world that wasn't there before. Aesthetics-first, control-forward — the goal is creative agency, not a roll of the dice.

Why we exist

I am because
we are.

Ubuntu is a Nguni Bantu philosophy — a person is a person through other people. Your humanity is bound up in everyone else's. It's been carried across generations of African thought as a way of organizing how people live, work, and decide. It isn't a tagline we picked out of a brand workshop. It's the principle we actually run the lab on.

Ubuntu reframes the question. Most product thinking starts with what does the user want? Ubuntu starts with who is the user a part of? That's not a small move. It changes who gets considered first. It changes what counts as success. It changes which trade-offs are acceptable when the easy ones cost someone you'll never meet.

Every product we build starts with the same question: who's underserved by the current version of this technology? The communities that need AI most — civic, creative, educational, the everyday people who never see themselves in a demo — are the ones we build for first. Not a phase-two roadmap item. The first cut. The shape of AI is being decided right now, and the rooms making those decisions don't look like the world the tools will land in. We don't think the answer is exclusion in reverse. We think the answer is showing up early and building thoughtfully.

That changes how we work, too. Free tiers aren't a marketing ploy here — they're a baseline. Open source where we can. Documentation we'd actually want to read. Support that treats people like they're owed an explanation, not a ticket number. It's a slower way to build, and a less marketable one. We measure by how long the work lasts, not how loudly it launches. The pace might disappoint a VC. This isn't their lab.

Edge in, not center out

The strongest tools come from the strangest places. Most AI right now is built for the same audience by the same people, and the result is a stack that's brilliant at one demo and useless to most of the world. We start at the margins on purpose. The communities that need AI most rarely have a seat at the table where AI is being designed. We build with them in mind first, because the work is sturdier when it has to hold up there before it ever needs to hold up anywhere else.

How we live this

Craft is respect

It's tempting to ship fast and iterate. We do the opposite. Every product we put our name on gets the same care we'd want if we were on the receiving end. RAG that actually retrieves what's relevant. Generation pipelines that don't fight the user. Support that explains the answer instead of escalating the ticket. Polish isn't the goal — the goal is that someone using our work feels like the people who built it took them seriously.

How we live this

Slow on purpose

AI moves fast. The work that lasts moves slow. We don't chase trends. We don't ship a thing because it'd be a good launch. We build things we want to live with, because the people using them will, too. The pace looks underpowered next to the news cycle and right next to a portfolio that's still standing in five years.

How we live this