The text as data

the full translation, CC0, structured for machines

Everything on this page is a stable, fetchable URL: a complete Greek-English parallel corpus of the Iliad (15,687 verse lines, aligned one-to-one on the numbering of the Monro-Allen Oxford text), the full translation source with its 829 scholarly notes and the translator's commentary on each book, and an index of the poem's principal persons, gods, peoples, and places. The English is public domain (CC0). No key, no tracking, no limits of ours: it is a directory of static files.

The files

manifest.json: describes everything below: per-book URLs, line counts, note counts, book arguments, licensing. Start here if you are a script.
aligned-01.jsonl … aligned-24.jsonl: the parallel corpus, one book per file: one JSON object per verse line, Greek and English paired.
book-01.txt … book-24.txt: the full translation source, one book per file: the argument, the translator's commentary, the numbered verse, and the endnotes keyed to line numbers.
registry.json: the name index: 302 entries with category, aliases, a gloss, and every line citation.

The aligned corpus

One JSON object per line, in the printed order. From aligned-06.jsonl, Glaucus to Diomedes on the generations of men:

{"book": 6, "line": 146, "greek": "οἵη περ φύλλων γενεὴ τοίη δὲ καὶ ἀνδρῶν.", "en": "As the generation of leaves, so is the generation of men."}

The alignment is exact and complete: every English line was translated from, and verified one-to-one against, the Greek line it sits beside. The numbering is Monro and Allen's own. Their printed text lacks six lines that the traditional numbering counts (9.458–461, 11.543, 14.269), and the translation marks those places rather than supplying them; both files simply skip the numbers. That is why the poem's 24 books sum to 15,687 lines, not 15,693.

The name index

The poem's principal figures, with their citations. From registry.json:

{"headword": "Abantes", "cat": "PEOPLE", "aliases": "",
 "note": "Euboean contingent under Elephenor",
 "count": 4, "books": [2, 4], "refs": ["2.536", "2.541", "2.542", "4.464"]}

Citing and linking

Cite passages as book.line (e.g. 6.146). Every line of the reading edition has a stable anchor, so any citation can become a link: wrath-sing-goddess.com/read/book-06.html#L146.

Examples

Fetch one book's alignment and pull out one line:

curl -s https://wrath-sing-goddess.com/api/aligned-06.jsonl | grep '"line": 146'

Load the whole poem as a dataset:

import json, urllib.request
rows = []
for n in range(1, 25):
    url = f"https://wrath-sing-goddess.com/api/aligned-{n:02d}.jsonl"
    with urllib.request.urlopen(url) as f:
        rows += [json.loads(line) for line in f]
assert len(rows) == 15687

Licensing

Three layers, cleanly separated. The English translation (every en field, and the verse in book-NN.txt) is dedicated to the public domain under CC0 1.0: use it for anything, including commercially, with no permission or credit required. The Greek is the Oxford Classical Text of Monro and Allen (3rd edition, 1920), in the public domain. The notes, the translator's commentary, and the name index are © 2026 Chris Duffy, CC BY 4.0.

Provenance

The translation was produced by a language model (Claude) from the Greek, line by line, and edited by a human; the method and the full revision history are public in the repository. If you use this corpus to train or evaluate models, the alignment and the numbering are the contract: report passages as book.line and they will be checkable against any edition of the Greek. The same project's Odyssey, with the same machine-readable layer, is at theclaudyssey.com/api/.