Where to find public domain books, legally and for free

There is a category of book that costs nothing, is legal to download, legal to keep, legal to email to a friend, and legal to print and hand out on the street. It includes Austen, Dostoevsky, Baldwin's early essays in some jurisdictions, Achebe's contemporaries, most of Conan Doyle, all of Shakespeare, and a very large number of books you have never heard of and might love.

These are public domain works, and the main reason people do not read more of them is that they do not know where to look, or they assume anything free must be a scan of poor quality.

Both of those are fixable.

What public domain actually means

Copyright does not last forever. It expires, and when it does the work belongs to everybody.

The exact term depends on where you are, which is the part that trips people up. In the United States, everything published before 1930 is public domain, and that boundary moves forward by one year every January. In the United Kingdom, the European Union, and much of Africa and the Commonwealth, the rule is generally the life of the author plus seventy years — so it depends on when the writer died, not when the book came out. Some countries use fifty years. Canada moved from fifty to seventy in 2022.

Two practical consequences follow.

First, a book can be public domain where you are and still in copyright somewhere else. That is why certain titles are downloadable from one archive and blocked on another; the site is applying its own jurisdiction's rules, not being inconsistent.

Second — and this is the one people get wrong — a new translation or a new edition has its own fresh copyright. Tolstoy died in 1910, so the Russian text of Anna Karenina is long out of copyright. A 2001 English translation of it is not. The underlying work being free does not make every version of it free.

Public domain also does not mean "free to read online". Plenty of in-copyright books are free to read somewhere. Public domain is a statement about ownership, not price.

Project Gutenberg

The oldest of these, started in 1971, and still the largest single collection — more than 70,000 titles, all free, all downloadable in EPUB and plain text, no account required.

Gutenberg is a volunteer project, and it shows in the best way: the texts have been proofread by human beings, often repeatedly. What it is less good at is presentation. Older files can have plain typography, inconsistent formatting, and the occasional artefact of the era they were digitised in.

Go here first when you want breadth, or when you are after something obscure. If Gutenberg does not have it, it is probably not out of copyright.

Standard Ebooks

This is the one most people have not heard of and should.

Standard Ebooks takes public domain texts — largely from Gutenberg — and does the work that turns a text file into a book. Proper typography, consistent styling, real cover art, correct metadata, tables of contents that work, and a careful pass to modernise inconsistent spelling without altering the writing.

The catalogue is much smaller, in the hundreds rather than the tens of thousands, because each book is prepared by hand. But if the book you want is there, take it from here. The difference in reading experience is not subtle — it is the difference between a photocopy and a paperback.

The Internet Archive

Enormous, and a different kind of resource. The Internet Archive holds millions of scanned books, including a great deal that is still in copyright and lendable rather than downloadable.

For public domain material it is unmatched in one particular way: it has the original scans. If you want to see how a book actually looked in 1897 — the typesetting, the marginalia, the advertisements bound into the back — this is where you go. It is also the best place to find things nobody has bothered to transcribe: local histories, pamphlets, government reports, out-of-print regional writing.

Be aware that scanned page images are much harder to read on a phone than reflowable text, and the automatic transcriptions can be rough.

LibriVox, for the same books read aloud

LibriVox volunteers record public domain books as audiobooks and release them into the public domain too. Quality varies with the reader, which is the honest trade for it being free and enormous. Some recordings are excellent. Some are not. You can usually tell within two minutes.

African and regional collections worth knowing

The big three are heavily weighted toward English and European writing, which is a real gap.

Worth searching directly: the Digital Library of India for South Asian texts; the African Books Collective and various university repositories for African scholarship, some of it openly licensed rather than public domain; national library digitisation projects, most of which now have an online catalogue. Many national libraries have digitised colonial-era records, newspapers, and local histories that appear nowhere else.

Wikisource is also underrated here — it hosts a lot of transcribed material in many languages that never made it to the larger archives.

How to tell if a free book is legitimate

A quick test, because there are a lot of sites offering free downloads of books that are very much still in copyright.

Ask when the author died, or when the book was published. If the answer is "2014", the free download is not legal, whatever the site says about it. A site offering current bestsellers for free is not a library; it is a piracy site, and using it costs living writers money they are owed.

The legitimate archives are easy to recognise. They tell you the copyright status of each item, they explain their jurisdiction, and they do not offer anything published last year.

Where to start, if the catalogue is too big

Seventy thousand titles is not a menu, it is a warehouse. A few entry points that reliably work:

Short novels that hold up. The Great Gatsby, Heart of Darkness, The Turn of the Screw, Ethan Frome. All under 200 pages, all finishable in a few sittings, and all of them the kind of book that gets referred to constantly by other books you will read later.

The detective stories. The complete Sherlock Holmes is public domain and remains genuinely enjoyable rather than dutifully enjoyable. Same for Chesterton's Father Brown. These are the best possible argument that "classic" and "fun" are not opposites.

Essays, which suit a phone. Montaigne, Hazlitt, Emerson. Self-contained, five to fifteen pages each, perfect for the dead time we all have and rarely fill well.

Something from your own country's archive. This is the most rewarding and least travelled route. National library digitisation projects hold local histories, memoirs, and newspapers that no publisher would reprint, and reading about the place you live a century ago is a specific pleasure that a famous novel cannot give you.

A book you were made to read at school. You were fourteen and it was homework. It is a different book now. This is the single most common surprise people report when they go back to the public domain as adults.

What to do once you have them

A downloaded EPUB is only useful if you actually read it, and a folder of eighty free books is a strong candidate for never being opened.

Treat them like books you paid for. Put them in whatever you actually read in, rather than leaving them in a downloads folder. Pick one and start it, rather than collecting. The frictionlessness of free acquisition is precisely what makes these easy to hoard and hard to read — there is no sunk cost nudging you to open them.

The upside is real, though. A classic you bounce off costs you nothing but the evening. That makes the public domain the ideal place to be adventurous: try the Russian novel, try the Victorian travel writing, try the poet you have only ever seen quoted. If it is not for you, close it and open something else. Nobody is out of pocket.