Not another agency. A subscription to ship.

€10,000 per month, per pod. Paid at the start, ex. VAT. Same number every month for as long as the pod is yours - no setup fee, no SOW, no overage clause buried in the small print.

Your roadmap can flex. The invoice can't.

Bank transfer. Invoice on signing, you pay it, the pod starts. Same the next month, and the one after. No deferred payments. Money in, work out.

There isn't one. Cancel on day three, the month refunds. Cancel on day twenty-nine, the month refunds. If we haven't earned the next month, we don't keep this one. Reputation lasts longer than an invoice.

Two documents: a mutual NDA and a contract. Both in English. The NDA can be yours or ours, as long as it's mutual and aligns with our model - the protection runs both ways.

The contract spells out everything - payment, refunds, scope, work, IP - in plain language. No addendums, no second contract showing up later.

Our contract and mutual NDA cover every engagement. If your jurisdiction or procurement chain needs extra paperwork - a DPA, a security questionnaire, an addendum from your security team - we'll cooperate, as long as it doesn't ask us to reshape the pod. We'll sign what your legal team needs. We won't rewrite how we work.

A pod is a unit of output, not a head-count. Senior engineers, heavy AI tooling, every day - sized to ship the work of two seniors running flat-out. We don't promise the number of people on the keyboard. We promise the velocity.

Not your problem. Continuity is our job. Every artifact - code, tests, docs, infra, the prompt library - is in your repo from day one. A new senior steps in and reads what's already there. The pod's velocity holds. We staffed it. We re-staff it.

Async, week to week. You send tasks in Slack - text, Loom, voice notes, whatever's quickest. We read them, ask questions if anything's ambiguous, and start work. Every seven days, a recorded walkthrough of what shipped that week lands in the channel.

If a task doesn't fit in the current month, it rolls into the next. If a task is bigger than one pod, we split it - across months, or across pods if you've reserved more than one. The queue is yours to reshape, ours to keep moving.

No standups, no status check-ins. The walkthrough is the status.

The engineers. There's no project manager, no account manager, no game of telephone. You message the people writing the code, and they reply directly. Decisions, questions, edge cases - same humans, same Slack.

When something goes sideways - and on any real engagement, eventually something will - we catch it early. A light AI pulse runs on our channels and flags the first signs of friction before they fester. If it doesn't sort itself out in writing, we book a call. The only meeting we'll put on your calendar is the one that actually moves something.

Sofia is GMT+2 in winter, GMT+3 in summer. Europe overlaps fully. The rest of the world overlaps partially - our morning or evening, depending on where you are.

The work runs async by design. Slack threads, Loom walkthroughs, voice notes when text won't do. The cadence is built for time zones, not against them. Calls happen when they'd actually move something forward, which is rarely. Longitude isn't the problem you think it is.

Either, and we don't have a preference. What changes between a blank repo and somebody else's codebase is the first week, not the answer.

Picking up an existing project, the first week is reading, not writing. The architecture, the commits, the parts of the system the team flinches at - usually a strong signal of where to start. Commits start the following week, on solid ground.

Greenfield is a luxury, not a prerequisite.

No. The pod doesn't slot in alongside an in-house team.

Engineering cultures don't blend. An async one and a calendar-driven one in the same repo means one drags the other down. We can't sign for code we didn't write or review. And two unrelated teams committing into the same codebase produce merge conflicts, duplicated work, and a steady drip of "why did you build that, we already had it" - problems with no clean fix and, mostly, no fix at all.

Pick a team. Either is better than both.

No. They're how we ship - the contract has a non-solicit for exactly that reason. The pod already is a senior engineer, long-term, without the recruiter or the hiring loop. Take the pod. That's what the model is for.

You do, from the first commit. The repo is yours, the IP in it is yours, no clause to negotiate. Code, tests, docs, infra, the prompt library we build on top - everything we produce gets committed where you can see it.

If you ever leave, you leave with a working system and the history that explains every line of it. Nothing parked in a private repo on our side.

No. Every model and tool we use runs on an enterprise tier that contractually excludes training on customer data. Repo access stays scoped to the engineers on your pod. Secrets stay in your vault. NDAs are signed before the first invite to the repo goes out.

We don't borrow your code. Not even by accident.

The model writes. The senior engineer edits, reviews, refactors, and signs the commit. Same way good code shipped in 2019, just faster.

The cliff we wrote about above? We don't drive off it.