Catalogue No. CH-2026.B An applied lab · a venture studio · est. 2026 Last revised — 22.V.2026
The studio at Clandestine Hours
Open to letters Say hello →
Entry № 000
Type
Applied lab / Venture studio
Domain
Agentic systems / Precision software
Status
Open to letters
Preface

We experiment on ourselves first. Then we ship, partner, or back.

A venture studio at the frontier of agentic systems. Our products are the experiments. The consulting funds the lab. The checks we write follow what we learned. The teaching closes the loop.

Key — ● open ● in motion ○ shipped
LIS · LOCAL
--:--
Working — across timezones
I. Four rooms

The studio, in four modes of work.

Each room funds the next.
The lab is the centre of gravity.
Room 01 Mode

Build

Own products. The lab's experiments.

Experiments, all of them — built because we needed them, shipped because they worked, iterated because real users revealed what mattered. Rarely killed. Each one teaches us something about how software should fit a person.

Commonplace · Nappy Notes · Book Time · fmaps · doctrine
Room 02 Mode

Partner

Full products. Agentic systems. Precision software.

We build the whole thing — not just automations. Full products, agentic systems, serious infrastructure. Scoped precisely to the real problem, sized to what it actually needs. We show up, build what matters, leave it running.

HVAC partnership · specialised coding agents
Room 03 Mode

Back

Sweat equity. Founding engineers as a service.

We embed as founding engineers or fractional technical leadership — hands-on, in the engine room, not the boardroom. Equity, tokens, or both. Capital too, when the fit is there.

Sweat equity · tokens · operator capital · by introduction
Room 04 Mode

Teach

Curated sessions & public talks.

Private sessions for teams crossing into agentic workflows. Lectures on what this moment means for how software gets built. Focused on the people doing the work, not the people managing it.

Adapted to your team · the working version
II. The Catalogue

The work, 2024 — present.

Name
Note
Kind
Since
In brief
Status
CH-P-01
Turn private knowledge into public signal.
Product
2025
Entering private beta. A quiet place to write in public.
In Motion
CH-P-02
Parenting is hard. Tracking shouldn't be.
Product
2024
Born from necessity, days before our first child.
Maintenance
CH-P-03
Tools for book lovers.
Product
2025.Q4
Fun tools for readers who can't decide what to read next.
Active
CH-X-01
Field Platform
Desktop & web suite for a Portuguese HVAC company.
Partnership
2026
Stock management, automated label printing, and custom factory tracking — built around how they actually work, not the other way around. Desktop and web, conflict-free.
Active
CH-X-02
MuleSoft Agents
Strategic advisory · proof of concept.
Advisory
2026
Feasibility case for specialised coding agents in MuleSoft workflows. Strategic advisory + POC — dashboard and a tuned harness to make the argument concrete. Delivered. Follow-up in the pipeline.
Delivered
CH-T-01
Codebase cartography for AI teams.
Tool
2026
Maps territories, dependencies, blast radius. Strategic governance for codebases where agents do the work.
Coming
CH-T-02
Product memory for AI coding harnesses.
Tool
2026
Gives agents structured product context before they code. Tracks decisions, hypotheses, and what was learned — so the next session isn't cold.
Coming
III. The longer arc

Most software will be ephemeral. That's fine.

The idea that software should last forever was always a cost of distribution, not a virtue. Agents change what's worth building — and for how long. We build full products and serious systems. But we build them precisely: scoped to the real problem, sized to what it actually needs, designed to earn their place. When something better exists, you move on. No legacy. No maintenance tax.

Every engagement teaches us something about that fit. The longer arc is a platform where it's native — where the software shapes itself to the person, not the other way around.

Building quietly. One step at a time.

If you want early access or to be a design partner — let us know.

The log — recent
2026
Every session starting cold is a solved problem. Built the memory layer so agents know what's being built, what was decided, and what the team learned — before they write a line.
2026
Agents were making confident, breaking changes to files they had no business touching. We built the map so they'd know the territory before acting.
2026
Some engagements stay off the record.
2025
Book Time launched — fun tools for readers who can't decide what to read next
2025
Not everything ships with a name attached.
2025
Nine years in, we scrapped everything and rebuilt. Same partnership, completely different software.
2025
The data model for how teams actually work is harder than it looks. Architecture phase begins.
2025
Strategic advisory and proof of concept for specialised coding agents in MuleSoft workflows. Delivered.
2025
After two to three years of brewing, Commonplace is ready for its first users. Invites going out this quarter.
IV. Correspondence

Write us a careful brief.

Prompts and specs are the craft, on both sides of a working relationship. The best first messages come from founders who've already pressure-tested their thinking.

Before you write — use these to get to version zero:

A prototype and a paragraph beats a deck. A sharper problem gets a sharper answer.

hello@clandestinehours.com →
brief.md — template plain text, please
# Who you are
[Name, role, company, one sentence on what you do.]

# What you're building
[The thing. What exists today, what's missing.]

# The hour you're short of
[The specific constraint or decision you're stuck on.]

# What "done" looks like
[Outcome, not deliverable. Six weeks, or tell us why not.]

# Constraints / non-negotiables
[Stack, timeline, budget signals.]

# How you found us
[Optional, but we're always curious.]