A sourced compendium
Angels, demons,
and where they came from.
Every figure here is traced to the text that names it, and most of them turn out to be older than the tradition that claims them — Canaanite, Mesopotamian, Egyptian, Persian, Greek and Norse gods, renamed and repurposed. Where the sources disagree, this site shows the disagreement instead of picking a winner.
- Entities
- 328
- Source texts
- 121
- Attestations
- 725
- Powers indexed
- 42
- Pantheons
- 15
- Syncretic chains
- 11
What is actually in here
When the names appear, and when the books do
The same strip that heads the timeline, on the same five eras and the same weighting, so following it through leads to the shape you just saw. Each column is an era, bar height is how many figures are first attested in that slice of it, and under it the count of texts surviving from the same stretch. The names arrive early and the books that catalogue them arrive late — which is the argument the rest of the site spends its time evidencing.
New here
Three ways in
Look something up
Find a figure
Every angel, demon and deity in the corpus, filterable by tradition, rank and how well attested it is. Start here if you arrived with a name in mind.
328 figures · 121 textsBe shown something
Read a study
Questions the corpus can settle by counting — how old the Goetic demons really are, when hell acquired a peerage, what the plagues of Exodus were aimed at. Start here if you want the interesting part first.
10 written · 725 attestations behind themUnderstand the method
See how it works
What an attestation score measures and what it deliberately does not, how sources are weighted, and where the corpus knows it is thin. Start here if you intend to rely on any of this.
The score, and what it will not tell youBefore you rely on any of it
How to read an entry
Four rules govern everything on this site. They are the difference between a compendium and a list of assertions, and they explain most of what looks unusual about the entries.
Every claim names its text
Nothing here is asserted on the site's own authority. A profile says what a named source says, with the date of that source attached, so a claim from a Bronze Age tablet is never silently mixed with one from a Victorian compendium.
Disagreements are kept
Where two texts contradict each other, both are shown. The Ars Goetia and Pseudomonarchia give different ordinals to the same spirits; the corpus records both rather than choosing, because the disagreement is itself the evidence.
Absence is recorded
“None” is an answer. Several figures have no cult site, no attested worship and no early source, and the entries say so plainly instead of leaving a blank that reads like an oversight.
Weak evidence is graded
Where a tradition rests a claim on a passage that will not carry it, the claim is recorded accurately and its textual basis is graded. The grading is about the strength of the evidence and never about whether anything is true.
Twenty minutes
A first path through
- 1
Read one profile
Lucifer is a good first entry because it is really two figures. Watch the name arrive in Isaiah as a title for a Babylonian king, and leave the Dictionnaire Infernal in 1863 as an emperor.
- 2
Look at what it is built from
Every claim on that page points at one of these texts. They run from Bronze Age tablets to Edwardian occult printing, and the date of a source is doing a great deal of work.
- 3
See two authorities contradict each other
The competing ranked orders, drawn as trees. They do not agree with one another, and no attempt is made here to make them.
- 4
Then read a study
Once the shape of the corpus is familiar, this is what it can be asked. Hell's peerage turns out to be datable, and much younger than the demons wearing it.
What the corpus answers
Latest studies
A study from the corpus
What the Goetia does not care about
It contributes nothing to the corpus's affliction claims, gives most spirits no harmful office at all, and connects none of them to each other. What it supplies without exception is a legion count.
The most famous demonology in English lists seventy-two spirits with ranks, legions and offices. What does it never say about any of them?
A study from the corpus
How many angels are actually named?
Two names are in books every canon receives. Raphael is deuterocanonical, Uriel is in an appendix nobody received, and fifteen of the entries people count as angels are plurals, ranks or offices rather than names.
Seven archangels with portfolios, nine choirs in three spheres, a name for every office. How much of that roster is in the books?
A study from the corpus
The divine council
Every Canaanite deity in the corpus is named somewhere in the Hebrew Bible. No Egyptian one is, out of forty-three — and the institution comes through with the vocabulary.
Psalm 82, Deuteronomy 32, 1 Kings 22 and Job 1 describe a heavenly assembly with members and business. How much of it is inherited?
Everything else
The rest of the corpus
Hierarchies
The competing ranked orders, drawn as trees.
Denominations
The churches that licensed or condemned all of it.
Syncretism
How one culture's god becomes the next one's demon.
Places
Where the cult actually was, mapped with coordinates.
Symbols
Seals, wards and divine names, dated to first use.
Afflictions
What each text says a spirit does to a person.
Rites
The exorcism liturgies, clause by clause.
Timeline
The source texts in order, tablet to compendium.
Traditions
The twenty-two traditions the material comes from.
Glossary
Decan, lamen, qliphoth — the vocabulary, defined.