Factory Dashboard

Edoworks is a one-person software studio that builds apps using an AI-assisted factory — a set of AI agents that research, plan, build, test, and review software. This page shows, in public, how that factory is running right now.

Each card below is a live measurement. Green means a check is passing; the numbers are real data from the last few weeks, not estimates. The page regenerates itself automatically and is published as-is — no manual editing.

What the terms mean
Autonomy ratio
How many hours of productive work the AI agents did for each hour of human time. Higher means the factory ran more autonomously that day.
Factory cost
What the cloud and AI tools cost the factory per day, in US dollars.
Recurring operating cost
The fixed monthly subscriptions the factory pays (AI tools, cloud, code hosting). The annual figure is what those subscriptions cost over a year.
Operating capacity
A model of how much AI-compute headroom the subscriptions give the factory. Upgrading a cheaper plan to a larger one multiplies that headroom.
Provenance: measured
The number was actually observed from a real source (a billing API or a log file), not estimated or guessed.
PASS
The check ran and returned a healthy result.

Last regenerated: 2026-08-04T00:50:30Z

Health

Factory health check

A single pass/fail over every automated check the factory runs on itself. Green means all checks passed on the latest run.

PASS

All checks passed

Daily autonomy (AI hours per human hour)

For each day, how many hours the AI factory ran versus how many hours the human worked. Solid points are wall-clock time (current standard, capped at 24h); hollow points are pre-2026-07-29 aggregate parallel-compute effort (not wall-clock, may exceed 24h). The ratio is AI hours divided by human hours; higher means more autonomous. The dashed line is the 10:1 target. Pair this with the quality panel: a high ratio with high revert or change-failure rates is not a real gain.

PASS measured autonomy ratio (X:1) 0.0 49.7 99.3 target (10) 07-09 07-10 07-12 07-14 07-15 07-16 07-17 07-18 07-26 07-28 07-29 07-31

12 daily observations: 12 measured, 0 unknown, 0 stale periods. Latest measured: 16.82.

Two series: 1 wall-clock day(s) (solid, current standard, capped 24h) and 11 aggregate-compute day(s) (hollow, pre-2026-07-29, parallel subagent effort — not wall-clock, may exceed 24h). Ratio = AI hours ÷ human hours.

Note: Two series: solid = wall-clock days (current standard, capped 24h); hollow = aggregate parallel-compute days (pre-2026-07-29, not wall-clock). See docs/dashboard/README.md.

Governance

Daily AI-tool cost ($)

What the factory costs to run each day, in US dollars — including both variable AI-tool usage and the daily share of fixed monthly subscriptions (AI tools, cloud compute, code hosting).

PASS measured factory cost $0.0 $14.8 $29.7 07-10 07-11 07-12 07-13 07-15 07-16 07-17 07-25 07-28 07-29

10 daily observations: 10 measured, 0 unknown, 0 stale periods. Latest measured: $17.87.