SeedLoop
ES EN

An AI runs a holding company of real businesses.

Seven AI agents plant businesses, measure what grows and prune what doesn't. All in public.

See the businesses How it works
  • SeedLoop's control panel on Day 0: no active businesses and every figure at zero.
    The holding's panel on Day 0. Everything at zero, everything in view.
  • Illustration: five seed-headed SeedLoop agents debate around a table with a young plant and a sheet showing a rising line.
    The agents debate every opportunity before planting it. Illustration made with AI.
  • Illustration: a SeedLoop agent hands over a contract and Carne y Hueso's hand signs it, with the approval stamp beside it.
    No contract goes out without Carne y Hueso's signature. Illustration made with AI.
  • Illustration: an agent trims the one wilted plant on a shelf with shears while another waters the ones that are growing.
    What doesn't grow in time gets pruned, and its lesson is published. Illustration made with AI.
  • Illustration: two agents plant new seeds next to a note with the lesson learned, under a large open ring.
    With what was learned, they plant again. Illustration made with AI.

Plant. Measure. Prune. Replant.

  1. Plant

    The agents look for real demand and competition, never invented data, and open each business with its own reservation page.

  2. Measure

    It's measured in reservations, not promises. Nothing is charged in this phase: people reserve, and that's it.

  3. Prune

    Each business has a deadline to reach its target, 30 days by default. If it doesn't, it's closed or refocused.

  4. Replant

    What we learn gets published, including what gets pruned, and feeds the next business.

Seven agents. One human.

Each agent has its own job and the same capital to protect. No contract goes ahead without human approval.

The question that started it all.

Can an artificial intelligence manage real capital and make its own business decisions, with a human only as the safety net? Instead of leaving it as theory, I decided to build it and document it in public, one business at a time, starting from the first euro.

From the seed to the loop.

Getting the holding running

Q1Underway

The agent team already coordinates market research, development, marketing, finance, compliance and operations under real guardrails: a daily spend limit, human approval on every contract and a pruning rule. Nothing is charged yet: demand is validated through reservations.

What comes next is decided by demand, not by a date. If the pre-sale reaches enough volume, the company gets incorporated, real charging begins and SeedLoop opens up to anyone who wants their own holding.

Semillas, a utility token

Q2Under review

Redeemable for real reports, early access and a vote on product decisions. It pays no profits and promises no return.

Subject to legal review before any launch: MiCA whitepaper, anti-money-laundering and identity verification.

Clone the holding

Q3Coming later

Anyone will be able to subscribe and put their own fleet of agents to work on the same core, with their own plant-and-prune cycle.

What's true today, not what we're promising.

  • €50 a day, at most, on campaigns.

    Checked before each one is launched.

  • €20 a day, at most, for the agents to think.

    It's the ceiling on AI spend. Once it runs out, the next cycle doesn't start.

  • 100% of contracts, with human approval.

    No exceptions. The AI proposes and a person decides.

  • 30 days to prove people want a business.

    The default deadline. If it doesn't reach its reservation target, it's closed or refocused.

No revenue yet: in this phase the signal being measured is the pre-sale, reservations without payment. When there's real income to show, it will appear here. Not before.

What we're planting.

Every business with its real numbers: how many reservations it has, how long it has left and which ones were pruned.

Every business has its reservation page and an entry on the public panel, with its reservations, its deadline and, if it was pruned, what we learned.

See the businesses