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 Odyssey (12,107 verse lines, aligned one-to-one on Murray's 1919 numbering), the full translation source with its 1,260 scholarly notes, and an index of every named person, god, people, and place in the poem. 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: numbered verse plus the scholarly endnotes (markdown footnotes keyed to line numbers).
registry.json: the name index: 434 entries with category, aliases, a gloss, and every line citation.

The aligned corpus

One JSON object per line, in Murray's printed order. From aligned-09.jsonl, the moment Odysseus tells the Cyclops his name:

{"book": 9, "line": 366, "greek": "Οὖτις ἐμοί γʼ ὄνομα· Οὖτιν δέ με κικλήσκουσι", "en": "Nobody is my name. Nobody is what they call me —"}

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 Murray's own, quirks included. His edition prints two pairs in transposed order (3.304–305 and 14.63–64; the files keep his printed order) and omits three lines as later interpolations (10.456, 16.101, 23.49; the numbering skips them). That is why the poem's 24 books sum to 12,107 lines, not 12,110.

The name index

Every named entity in the poem, with its citations. From registry.json:

{"headword": "Achaeans", "cat": "PEOPLE", "aliases": "the Argives; the Danaans",
 "note": "the Greeks — the poem's default name for the host",
 "count": 130, "refs": ["1.90", "1.239", …]}

Citing and linking

Cite passages as book.line (e.g. 9.366). Every line of the reading edition has a stable anchor, so any citation can become a link: theclaudyssey.com/read/book-09.html#L366.

Examples

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

curl -s https://theclaudyssey.com/api/aligned-09.jsonl | grep '"line": 366'

Load the whole poem as a dataset:

import json, urllib.request
rows = []
for n in range(1, 25):
    url = f"https://theclaudyssey.com/api/aligned-{n:02d}.jsonl"
    with urllib.request.urlopen(url) as f:
        rows += [json.loads(line) for line in f]
assert len(rows) == 12107

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 Murray's 1919 text (public domain), as digitized by the Perseus Digital Library (CC BY-SA 4.0; keep their attribution if you redistribute the Greek fields). The notes 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, the full revision history, and a measured analysis of its independence from prior translations are public: the repository and the independence page. 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.