Why Kaelum exists

Built to remember what markets make everyone forget.

Kaelum is a private research instrument built by one operator — to replace opinion with evidence, and memory with an archive.

Behind it: a discretionary trader for four-plus years. In early 2026, the potential of modern AI became impossible to ignore — and the decision was made to turn years of trading and market experience into something that outlasts opinion: a quant research system that remembers, measures, and explains itself.

Evidence over opinion

A view that isn't recorded, timestamped, and reviewed later is just a mood.

Memory over momentum

The archive keeps what the system knew at the time — especially when it was wrong.

Boundaries over noise

The public window shows posture and scale. The working room stays closed.

The story so far

Six phases, one instrument.

Every phase left something you can still inspect — the panel beside each one is live.

  1. AI & dev tooling$400
    Market data feeds$227
    Hosting & infrastructure$20

    the first invoices — from the public burn ledger, February

  2. 3,335records · append-only · nothing rewritten

    the archive counter — live

  3. GCNQESBTCCL
    27,612 source items scanned

    the field of view — live freshness

  4. 08910110476850
    07-05recorded decisions per day07-11

    recorded decisions per day — live

  5. 346merged changes · 9 weeks · since mid-May 2026
    9238958542231746
    May 11Jul 6

    merged changes per week — current codebase, since mid-May 2026

Today

The story continues in the archive.

One operator, one instrument — and an archive that keeps the score.

How the machine works →Open analytics →