Transparency by default
AI consulting firms keep their methods private. We expose ours in real time. If you see what we consume and produce, you know who you're talking to before even booking the diagnostic.
01
Open laboratory
Bubble's lab, in real numbers and real time.
We build AI agents, run them on our own operations, and share everything. Here's what's running here right now.
Live counters
What we consume, what's running, what we publish. Real, aggregated numbers. No internal names exposed.
—
AI tokens consumed this month
—
active routines
—
public skills
—
agents in production
—
yrs finance XP
—
code shared
Live, in the lab
Latest public outputs from our agents and the team: Substack articles, GitHub commits, published skills. Filtered to exclude automatic technical commits.
- Loading recent activity…
Why we expose all of this
Proof through doing
The 22+ agents running here are the same kind of thing we install at our clients': daily automation, briefings, audits, drafting. We don't sell a promise, we sell what we use every day.
For agents, not just humans
This page is also structured to be readable by ChatGPT, Claude, Perplexity, and other LLMs. The
/api/labs/stats.json endpoint is public. Our llms.txt lists all the
machine-friendly context.
First contact
Want us to install the same kind of lab at your place?
3 sessions, €2,000 excl. VAT, we install and we train. First call is free, 30 minutes, no commitment.
Book a free diagnostic (30 min)