About Storgy
A reading room for poetry, built in public, free to use.
A public-domain library of 268 poets and 2,422 poems, an 94-theme atlas connecting them, and five free tools that read, write, check, and grade poetry alongside you.
The library
Free- Poets
- 268
- Poems
- 2,422
- Themes
- 94
- Tools
- 5
- Public-domain corpus
- No ads, no signup wall
- Two AI runs per day, free
What Storgy is
A poetry knowledge engine, not a magazine.
Three things sit underneath this site. A catalogue of poets. The poems they wrote. And a curated atlas of themes — wedding readings, funeral readings, identity, sleep, the sea, despair — that runs across both.
Each layer is built to be browsed by a real reader with a specific reason for being here. A student with an essay due Friday gets a clear read on what the poem is doing. Someone choosing a wedding reading gets a short, tasteful poem they can read aloud without losing the room. A returning reader gets adjacent works they haven't read yet.
Principles
Six rules the site is built on.
Public-domain first
Every poem on Storgy is sourced from public-domain corpora and cross-checked before it is published. Project Gutenberg is the primary text source; Wikidata cross-checks author and work metadata.
An index, not a feed
The library is arranged for browsing in three directions — by poet, by poem, by theme — and the entries connect across all of them. No algorithmic ranking, no infinite scroll.
Schema-constrained AI
Every analysis comes from Claude Sonnet asked to return a strict JSON shape, so each result carries the same seven sections. The model is constrained by schema, not free-form prose.
A second pass for tone
Every AI-drafted block runs through a separate humaniser pass that strips boilerplate — stacked em dashes, inflated symbolism, the words iconic, timeless, masterpiece.
Living poets, bio only
Storgy does not publish full texts by living poets. Their pages carry a biography, a list of representative works, and direct links to buy the books from publishers and bookshops.
No ads, no upsell wall
Five free tools share a two-runs-per-day cap across the whole toolkit. No advertising, no account required, no paid tier between a reader and a poem.
Where the poems come from
The corpus pipeline.
Public-domain status is verified per poem before it ships. The pipeline runs on three rules, and every entry on the site carries a provenance record that links back to its source.
The full sourcing methodology, including verification dates and per-poem provenance records, lives on the sources page.
Read the sourcing methodologyPublic-domain first
Project Gutenberg is the primary text source. Wikidata cross-checks author and work metadata. Poets.org's public-domain section fills gaps.
US-only geo-gate
US copyright rules give the largest pool of free text. Visitors from the US (and search crawlers) see full bodies. Visitors from the UK and EU see authorised excerpts where the work is not yet in their local public domain.
Living poets, bio only
No full texts by living poets. Their pages carry a biography, a list of representative works, and direct links to buy the books.
Who runs it
A small editorial team behind the index.
Storgy is built by a small editorial-and-engineering team working in public. The corpus pipeline, the toolkit, the editorial pages, and the design all run on the same engineering you see in the margins of this site.
Because the project runs lean, the timeline is deliberate. Accuracy and per-page provenance come ahead of raw volume. Attribution corrections are deployed in the same week they are reported, and the team takes public responsibility for the errors it does ship.
The aim is short, concrete prose that reads like a working teacher's notes — not an essay padded out to look smart. A poem is the work; the tool is the marginalia. The reader still does the reading.
What's next
On the editorial roadmap.
The corpus grows every week. These are the threads being worked on now, in rough order of when they ship.
In progress · 2026
Eduqas 2025–2027 anthology
The 15 new anthology poems are being added in the order Founding Teachers vote for them, ahead of first assessment in summer 2027.
In progress · 2026
Comparison atlas
Paired-poem essays extending past the current set — two poems side by side, four reading axes, what they share, where they diverge.
Next up · 2026
Movements and translations
Editorial intros to Modernism, the Beats, and the Harlem Renaissance, plus a bilingual reader pattern for Cavafy, Lorca, and others.
Questions
Things readers and teachers ask about Storgy.
Why call it Storgy?
Storgē is the ancient Greek word for the steady, familial kind of love — the one that grows in everyday closeness rather than romance. It is the register most poetry actually lives in: attention, return, careful reading.
Is the AI work disclosed on every page?
Yes. Every AI-drafted block runs through a second humaniser pass before publishing, and every poem page links back to the public-domain source it was rendered from. The full provenance methodology lives at /sources/.
Why public-domain only?
Modern copyright law makes the difference between a free, comprehensive corpus and one that needs licensing deals for every estate. Public-domain texts are the largest body of poetry we can index, annotate, and serve without permission, and they cover most of the canon that schools and curious readers actually want to read.
Do you publish anything by living poets?
Not full texts. Living poets get a biography, a list of representative works, and direct links to buy the books from publishers and bookshops. The full-text reading happens at the publisher's site or in print, where it should.
How are the AI tools paid for?
Two runs per day per visitor, shared across the five tools, free for everyone. Premium subscriptions (Student and Teacher Pro) pay for the API spend; the free tier covers the public corpus on its own. The pricing page at /pricing/ has the full breakdown.
Can I report an error or suggest a poet?
Yes — write to hello@storgy.com. Attribution corrections, missing poets, line-break issues, and new theme suggestions are all welcome. Corrections typically ship in the same week.