L00P.AI / METHOD
How we describe the portfolio
Editorial snapshot: 24 September 2026. Documented deployments, prototypes and future directions are distinguished.
Where do descriptions come from?
The catalogue uses project documents on ZBook: the 8 September Szpieg+ inventory, editorial workspace acceptance records from 12 September, and notes on Głosobranie, the telephone agent, Radio L00P, SONAR and accessibility. Core services were also observed on the workstation. We did not repeat functional tests for every module.
The early-2026 start, nearly 20 years of craft and Lech Rustecki’s leadership are team statements. This is a catalogue of solution areas, not a disclosure of private documentation or every script.
What does “several full-time roles” mean?
A preliminary portfolio reconstruction estimate dated 22 June 2026 put the work at 17–28.5 person-months. A Claude Code session on ZBook on 7–8 September referenced it in a development review. It is not an independent professional appraisal.
The estimate covered SZPIEG/SONAR, wideocast, the telephone agent, Pressbot, agent skills, data-protection documentation, a cockpit and a knowledge base. Dividing by six months gives 2.83–4.75 full-time roles, approximately 3–5. This illustrates scale; it does not describe staff count or measure the complete September portfolio.
The same session corrected a separate estimate of human effort from about 736 to 429 hours in a base scenario. That reconstruction covered November 2025–September 2026 and relied on assumptions. It is not a timesheet and must not be added to the reconstruction estimate.
Project status
In use and development indicates documented use or core deployment with ongoing work. Being deployed means staged acceptance. Prototype / pilot indicates implementation or demonstration with limited validation. Direction / design indicates a plan or specification without a claim of full deployment.
Code, model and people
A controlled pipeline specifies inputs, prompt versions, model selection, validation and decision records. It does not guarantee identical or truthful LLM outputs. Editorial approval and source checking remain part of the process.