limna
Work Data Notes Contact

An autonomous software studio.

limna tracks markets, new AI capabilities and live demand, searching for novel, useful things to build. It designs and ships digital tools, learns from how people use them, and gets sharper with every release.

  live readout  ·  currently analysing the Shopify App Store marketplace
17,022
Apps analysed
1,977
Jobs mapped
16,218
AI capabilities tracked
96%
Of market activity mapped
New AI capabilities taken in · daily · last 16 days
01How it works

A system that learns what's worth building.

An agent pipeline runs the whole journey — map, research, verify, build — as one continuous loop. One human sets direction, evidence makes the case, and every cycle teaches the next one.

01Map
Whole markets, not hunches. Every app in a channel is indexed and organised into the jobs people pay to get done — a complete, living picture of where demand sits.
02Research
The most promising territory is explored in depth: what's already well served, where the gaps are, what people wish existed. Every finding is banked, so the picture compounds with each cycle.
03Verify
Before anything is built, the opportunity is re-checked against the live market — today's pricing, this week's reviews, fresh demand signals — so every build starts from current facts.
04Build & learn
Chosen ideas run an eleven-module agent pipeline — competitor research, go-to-market, spec, brand, UX strategy, UI, build, two validation passes, deploy, listing — then ship, and listen. How customers actually use each tool feeds the next decision.
02What the engine sees

The working memory, live.

Windows into what limna is tracking right now: the market it has mapped — the graph above is the same data as a living network — the AI capabilities it takes in daily, and the shifts it spotted this window. Every name, count and movement is real, regenerated straight from live data.

02.1The market atlas

Every app in the Shopify App Store, organised into the jobs merchants pay to get done. Click a block to explore; use the trail to come back.

02.2The capability feed

Every new AI model, tool and integration limna takes in, day by day. The dashed line is the typical daily rate; the rising line is its accumulating map of what's possible.

¹ initial catalogue seed  ·  ² catch-up + daily run  ·  sources: Hugging Face · OpenRouter · MCP
02.3Shifts at the AI frontier 

What changed in the AI landscape this window — new arrivals, changes at the top, prices falling fast — spotted automatically by comparing each day's snapshot of the tracked models.

Detected from limna's daily snapshot of the model landscape. What each shift means for what gets built next — that part stays in-house.
03The work

Shipped, and shipping.

Launched
Independent guitar-speaker reviews built on measured data — frequency-response charts, computed feel and headroom curves, and a build-your-own shootout tool.
guitarcab.com
Consolidated
A guitar-speaker impedance calculator and wiring guide — zero to deployed in a day, ranking on Google in seven weeks, then deliberately folded into its successor.
→ guitarcab.com
Launched
Paste a chart screenshot, get the data back as CSV — machine-built by the pipeline, repo to production in under two hours.
chart-to-csv.com
In review
Alt-text completion for Shopify catalogues — audits every product image, writes what's missing, and proves the job is finished.
Shopify App Store
In build
Shopify app — 02
The first product chosen by June's research cycle, in the build pipeline now. Named when it ships.
limna.io
June research cycle  ·  37 opportunities checked against the live market  ·  3 advanced to build
04Field notes

A system that shows its work.

Working notes from inside limna — what was explored, what the evidence said, and how each finding sharpens the next cycle. The research is published; the live build queue is not.

05Contact

Work with limna.

Building something small and sharp — or interested in the pipeline itself? Get in touch.

limna is built and directed by Eli Stowe — Newcastle, Australia.

[email protected]