The Pāli Canon

The Pāli Canon in three forms, from one aligned source: the Pāli on its own, an English translation on its own, and the two side by side.

The Pāli has long been available digitally, but not really referenceable — you cannot find a passage without typing diacritics most keyboards do not have, and you cannot easily see which English sentence answers to which Pāli one. This is an attempt to fix both. Search here ignores diacritics entirely: type panna and you will find paññā.

The English is meant to stand on its own for anyone who simply wants to read it, and to link back to the Pāli for anyone who wants to check it. Where a text has several translations, they coexist rather than compete — no one rendering is treated as the correct one.

Pāli — the Mahāsaṅgīti edition, public domain. Use it to read or quote the source, or to check a term.

English — a continuous reading text, with nothing abridged. Use it if you just want to read the discourse.

Side by side — the two aligned paragraph by paragraph, and closer than that wherever the Buddha or an arahant is speaking. Narrative framing is aligned loosely, because only its boundaries matter; the doctrine is aligned tightly, because that is where the wording is worth comparing.

The Pāli root text is public domain and is offered for every sutta here. English translations are not all free to republish. Where a translation is still in copyright, the entry is listed with a 🔒 and the reason, and only the Pāli view is available. Nothing is hidden silently — you can always see what exists and why it is closed.

Peyyāla — the manuscript convention for “and so on”. A passage is written out in full once, and every later recurrence is abbreviated. It is why the Pāli often looks much shorter than the English: the translator wrote the repetition out, the Pāli edition did not.

Where the omitted text can be recovered, it is shown with the passage it came from. Recovered text is always marked as such and never merged silently into the source.

Two conventions for the same sound are in circulation — ṁ and ṃ — and different sources use different ones. Within a text the spelling is kept consistent. Diacritics that were lost getting the text into digital form have been restored only where a dictionary lookup settles the word unambiguously; anything genuinely ambiguous is left alone.

Reasonably, but check it where it matters. It was measured against segment-keyed data where the correct answer is known by construction: it recovers about 89% of pairings with at least partial overlap and about 78% exactly, against 35% and 26% for naive line-by-line pairing. It is a reading aid, not an authority — where a point turns on the exact wording, read the Pāli.