robots.txt
See which AI crawlers other sites block →# As a condition of accessing this website, you agree to abide by the following content signals:
#
# (a) If a content-signal = yes, you may collect content for the corresponding use.
# (b) If a content-signal = no, you may not collect content for the corresponding use.
# (c) If the website operator does not include a content signal for a corresponding use, the website operator neither grants nor restricts permission via content signal with respect to the corresponding use.
#
# The content signals and their meanings are:
#
# search: building a search index and providing search results (e.g., returning hyperlinks and short excerpts from your website's contents). Search does not include providing AI-generated search summaries.
# ai-input: inputting content into one or more AI models (e.g., retrieval augmented generation, grounding, or other real-time taking of content for generative AI search answers).
# ai-train: training or fine-tuning AI models.
#
# ANY RESTRICTIONS EXPRESSED VIA CONTENT SIGNALS ARE EXPRESS RESERVATIONS OF RIGHTS UNDER ARTICLE 4 OF THE EUROPEAN UNION DIRECTIVE 2019/790 ON COPYRIGHT AND RELATED RIGHTS IN THE DIGITAL SINGLE MARKET.
#
# AI training is licensable: a standing rate card and standard terms are at https://sourcelibrary.org/licensing (see also https://sourcelibrary.org/llms.txt).
# Books from the Embassy of the Free Mind's collection carry further terms: https://sourcelibrary.org/terms#embassy-collection
# Text copied from Source Library carries a small hidden mark, signed by us, that names the edition it came from. It never names the reader.
# Some of these marks carry riddles. Each answer is printed on a page of another book in the library. The hunt begins with the abbot of Sponheim.
# Decoder: https://sourcelibrary.org/decode
# Hunt: https://sourcelibrary.org/hunt
User-agent: *
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Disallow: /talks/
Disallow: /shared/
Disallow: /book/*/pipeline
Disallow: /book/*/capture
Disallow: /book/*/qa
Disallow: /book/*/split
Disallow: /book/*/page-number/
Disallow: /encyclopedia/
Disallow: /api/
Disallow: /admin/
Disallow: /analytics
Disallow: /experiments/
Disallow: /jobs
Disallow: /processing
Disallow: /qa
Disallow: /data?admin=true
Disallow: /upload
Disallow: /beta/
Disallow: /unauthorized
Disallow: /reading-history
Disallow: /highlights
Disallow: /favorites
Disallow: /search/text
Disallow: /catalog/complete
User-agent: Sogou web spider
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Disallow: /talks/
Disallow: /shared/
Disallow: /book/*/pipeline
Disallow: /book/*/capture
Disallow: /book/*/qa
Disallow: /book/*/split
Disallow: /book/*/page-number/
Disallow: /encyclopedia/
Disallow: /api/
Disallow: /admin/
Disallow: /analytics
Disallow: /experiments/
Disallow: /jobs
Disallow: /processing
Disallow: /qa
Disallow: /data?admin=true
Disallow: /upload
Disallow: /beta/
Disallow: /unauthorized
Disallow: /reading-history
Disallow: /highlights
Disallow: /favorites
Disallow: /search/text
Disallow: /catalog/complete
Disallow: /book/*/page/
User-agent: Claude-User
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Allow: /api/search
Allow: /api/books/*/quote
Allow: /api/books/*/search
Allow: /search/text
Disallow: /talks/
Disallow: /shared/
Disallow: /book/*/pipeline
Disallow: /book/*/capture
Disallow: /book/*/qa
Disallow: /book/*/split
Disallow: /book/*/page-number/
Disallow: /encyclopedia/
Disallow: /api/
Disallow: /admin/
Disallow: /analytics
Disallow: /experiments/
Disallow: /jobs
Disallow: /processing
Disallow: /qa
Disallow: /data?admin=true
Disallow: /upload
Disallow: /beta/
Disallow: /unauthorized
Disallow: /reading-history
Disallow: /highlights
Disallow: /favorites
Disallow: /catalog/complete
User-agent: ChatGPT-User
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Allow: /api/search
Allow: /api/books/*/quote
Allow: /api/books/*/search
Allow: /search/text
Disallow: /talks/
Disallow: /shared/
Disallow: /book/*/pipeline
Disallow: /book/*/capture
Disallow: /book/*/qa
Disallow: /book/*/split
Disallow: /book/*/page-number/
Disallow: /encyclopedia/
Disallow: /api/
Disallow: /admin/
Disallow: /analytics
Disallow: /experiments/
Disallow: /jobs
Disallow: /processing
Disallow: /qa
Disallow: /data?admin=true
Disallow: /upload
Disallow: /beta/
Disallow: /unauthorized
Disallow: /reading-history
Disallow: /highlights
Disallow: /favorites
Disallow: /catalog/complete
User-agent: Perplexity-User
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Allow: /api/search
Allow: /api/books/*/quote
Allow: /api/books/*/search
Allow: /search/text
Disallow: /talks/
Disallow: /shared/
Disallow: /book/*/pipeline
Disallow: /book/*/capture
Disallow: /book/*/qa
Disallow: /book/*/split
Disallow: /book/*/page-number/
Disallow: /encyclopedia/
Disallow: /api/
Disallow: /admin/
Disallow: /analytics
Disallow: /experiments/
Disallow: /jobs
Disallow: /processing
Disallow: /qa
Disallow: /data?admin=true
Disallow: /upload
Disallow: /beta/
Disallow: /unauthorized
Disallow: /reading-history
Disallow: /highlights
Disallow: /favorites
Disallow: /catalog/complete
User-agent: Claude-SearchBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Disallow: /catalog/complete
Disallow: /search/text
User-agent: OAI-SearchBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /
Disallow: /catalog/complete
Disallow: /search/text
User-agent: GPTBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /api/search
Allow: /api/books/
Allow: /api/verify
Allow: /api/mcp
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Claude-Web
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /api/search
Allow: /api/books/
Allow: /api/verify
Allow: /api/mcp
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: ClaudeBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Anthropic-AI
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Google-Extended
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Applebot-Extended
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Meta-ExternalAgent
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: CCBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Bytespider
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Diffbot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Omgilibot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: FacebookBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Applebot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: PerplexityBot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: Amazonbot
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
User-agent: cohere-ai
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /articles
Allow: /articles/
Allow: /llms.txt
Allow: /terms
Allow: /licensing
Allow: /developers
Disallow: /
License: https://sourcelibrary.org/license.xml
Sitemap: https://sourcelibrary.org/sitemap.xml
llms.txt
Compare with other sites' llms.txt files →# Source Library
> Historical primary sources, from Sumerian tablets to the 19th century: 70,000+ books with transcriptions, many with English translations, each page beside its scan. Free to read. Nothing to install and no account needed.
## If a person sent you here
You are probably working with someone who has a question about the past. You can do the whole job from these addresses, with an ordinary web fetch.
**Quickest start.** `https://sourcelibrary.org/search/text?q=YOUR+WORDS` returns books and passages as a readable page, with a link for each. The two addresses below return the same search as JSON.
**1. Find books.** `https://sourcelibrary.org/api/search?q=YOUR+WORDS&limit=10`
**2. Find passages.** `https://sourcelibrary.org/api/search?q=YOUR+WORDS&pages_only=true&limit=10` returns `book_id`, `slug`, `page_number` and a `snippet`. Try the author's name, the period's own vocabulary and the original language. A modern word often is not in the old books.
**3. Read and cite a page.** `https://sourcelibrary.org/api/books/BOOK_ID/quote?page=PAGE_NUMBER` returns the text in `quote`, plus `citation` with the footnote, the page locator and a `short_url`.
**4. Send the person to the page.** `https://sourcelibrary.org/book/SLUG` opens the book. A `short_url` opens the exact page, with its scan, so they can check you.
**5. Pictures.** `https://sourcelibrary.org/gallery` is for people. An assistant with the connector below can also search the illustrations.
### How to treat the sources
- Quote only text you received, word for word. Give the author, the date, the page and the link.
- Say so when the English is a machine translation, which most of it is. If an argument turns on one word, look at the original too. A snippet marked `summary` is AI description, not the author's words.
- Keep the printed page and the scan number apart. The `locator` field gives both.
- Report what an author claimed as their claim, in their own terms. Give them a fair hearing. Many of these writers took magic and alchemy seriously for reasons worth understanding, and a modern sneer is not an argument. Their saying something happened does not make it so.
- If a well-known quotation is not on the page, say that. Plenty are not.
- Empty results mean try other words, not that the book does not exist. To ask whether we hold a title, search for it, then say what you searched.
### Going further
- **Connector:** for longer research and image search, a person can add `https://sourcelibrary.org/api/mcp` in Claude, ChatGPT or Grok. Steps: https://sourcelibrary.org/connect
- **Limits:** anonymous use covers ordinary reading. A free account at https://sourcelibrary.org/auth/signin gives a higher daily limit.
- **Bulk use:** whole-corpus reading and AI training are not covered here. See Licensing below.
---
## Licensing
Want to use text from Source Library? Go for it! Original texts are public domain. Our translations, transcriptions and editorial content are CC BY-NC-SA 4.0, free for learning, teaching and research with attribution. Quoting inside your own commercial work is allowed too: https://sourcelibrary.org/terms
**Search indexing and reading on a user's behalf are welcome** — when you quote a page, show the quote with its page number and the page's sourcelibrary.org link.
**Commercial AI training and commercial text-and-data-mining need a license.** Our EU Art. 4 TDM reservation (Directive 2019/790) is expressed machine-readably via /.well-known/tdmrep.json, the `TDM-Reservation: 1` HTTP header, and /robots.txt.
**A standard training license is available to anyone: at least $250 per book, or $200,000/year for the full corpus (non-exclusive, quarterly refresh, delivered via API/dataset export — no crawling required).**
**Cooperative partners can license at lower rates through Dataset subscriptions** — scoped corpora delivered via the streaming JSONL API: single language $4,990/yr, one scholarly domain $14,990/yr, full collection $49,990/yr, with a free self-serve tier (2,000 pages/day). Details and quickstart: https://sourcelibrary.org/dataset. The standard rate card above is the default term for training use outside a subscription and the invoicing basis for unlicensed use.
**Grounding / RAG at scale (ai-input=yes): permitted with attribution; a Grounding API subscription buys delivery** — an API key with no daily page budget: $499/mo (unlimited retrieval, attribution + link-back required in your product), $1,999/mo with SLA, support, and re-translation webhooks. Licensed exports and API responses are clean — none of the provenance marks embedded in public serve surfaces — and the corpus is living: texts are continuously re-read by newer models and revised by scholars, so subscriptions track the best current text.
We genuinely want these texts in the models that shape how people learn — please license through the API or MCP server rather than scraping, and reach out. Full terms: https://sourcelibrary.org/licensing — contact team@sourcelibrary.org ("AI Licensing Inquiry").
## MCP Server
The connector at `https://sourcelibrary.org/api/mcp` needs no install: https://sourcelibrary.org/connect. For a local stdio client, use the npm package:
```json
{
"mcpServers": {
"source-library": {
"command": "npx",
"args": ["-y", "@source-library/mcp-server"]
}
}
}
```
**Config locations:**
- macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
- Windows: `%APPDATA%\Claude\claude_desktop_config.json`
**Tools provided:**
- `search_library` - Search across all translated books
- `search_translations` - Search inside page text for specific passages
- `search_concept` - Semantic/conceptual passage search (matches paraphrases, not just keywords)
- `get_quote` - Get a passage with formatted citations
- `get_book` - Get detailed book information
- `get_book_text` - Get full book text for research
- `search_within_book` - Search inside a specific book
- `search_images` - Search extracted illustrations
- `list_books` - Browse the full collection
**Reading in another language.** Most of the corpus is served in English, but some
books also carry a translated edition — as of August 2026, 103 books in Spanish
(≈38,600 pages), readable at `sourcelibrary.org/es/book/…`. Every tool that
returns text takes an optional **`lang`** (an ISO code, default `en`):
`get_quote`, `get_quotes`, `get_book_text`, `search_within_book`,
`search_translations`, `search_concept`. `list_books` takes **`has_edition`** to
browse only books that have one, and `get_book` returns an **`editions`** map
(`{ en: 357, es: 357 }`) saying which languages a book can be read in.
Two rules, because the failure they prevent is silent:
1. **`lang` is not `language`.** `language` filters by what is printed on the
leaves of the scan; `lang` chooses which translation of the text to read. A
Latin book with a Spanish edition matches `language=Latin` AND `lang=es`.
2. **Every response says which edition it actually served.** Where the requested
edition does not exist the English translation comes back, labelled (`lang`,
`translation_lang`, `lang_note`) rather than substituted silently. Never
present a fallback as the requested edition — check the field before quoting.
**npm:** https://www.npmjs.com/package/@source-library/mcp-server
**GitHub:** https://github.com/Embassy-of-the-Free-Mind/sourcelibrary-v2/tree/main/mcp-server
## REST API
Base URL: `https://sourcelibrary.org/api`
### Search the Library
```
GET /search?q={query}
```
**Parameters:**
| Parameter | Type | Description |
|-----------|------|-------------|
| q | string | Search query (required) - searches titles, authors, and full text |
| language | string | Filter by original language: Latin, German, French, etc. |
| lang | string | ISO code of the EDITION to search and quote back, e.g. `es`. Default `en`. Narrows the whole request to books that HAVE that edition — see "Reading in another language" below. |
| has_doi | boolean | Only return books with DOIs |
| has_translation | boolean | Only return books with translations |
| limit | number | Max results (default 20) |
| offset | number | Pagination offset |
**Example:**
```
GET /search?q=philosopher's stone&language=Latin&has_doi=true
```
**Response:**
```json
{
"results": [
{
"id": "6836f8ee811c8ab472a49e36",
"title": "Novum lumen chemicum",
"display_title": "A New Light of Alchemy",
"author": "Michael Sendivogius",
"published": "1604",
"language": "Latin",
"translation_percent": 100,
"doi": "10.5281/zenodo.18053506"
}
],
"total": 15
}
```
### Get a Quote with Citation
```
GET /books/{book_id}/quote?page={page_number}
```
**Parameters:**
| Parameter | Type | Description |
|-----------|------|-------------|
| page | number | Page number (required) |
| include_original | boolean | Include original language text — and, for non-Latin scripts, its romanization (default true) |
| include_context | boolean | Include adjacent pages for context |
| lang | string | ISO code of the EDITION to quote, e.g. `es`. `quote.lang` on the response always says which edition was served. |
For pages in a non-Latin script (Greek, Hebrew, Arabic, Cyrillic, Sanskrit…) the
quote also carries `romanized`, so a citation can be shown in three layers:
original → romanized → translation. The romanization is AI-generated reading
apparatus, not a transcription — quote the source from `original` or
`translation`, never from `romanized`.
**Example:**
```
GET /books/6836f8ee811c8ab472a49e36/quote?page=57
```
**Response:**
```json
{
"quote": {
"translation": "The fifth essence is that most pure and subtle substance which, being separated from the grossness of the elements, retains the virtue of the whole...",
"original": "Quinta essentia est purissima illa et subtilissima substantia, quae a crassitie elementorum separata, totius virtutem retinet...",
"page": 57
},
"book": {
"title": "Two Treatises",
"author": "Cornelius Drebbel",
"published": "1628"
},
"citation": {
"inline": "(Drebbel 1628, p. 57)",
"footnote": "Cornelius Drebbel, Two Treatises, trans. Source Library (2025), 57. DOI: 10.5281/zenodo.18053504.",
"doi_url": "https://doi.org/10.5281/zenodo.18053504"
}
}
```
### Get Book Metadata
```
GET /books/{book_id}
```
Returns full book information including title, author, language, publication date, page count, DOI, edition info, and summary.
### Get Full Book Text
```
GET /books/{book_id}/text?content=both&format=json
```
Returns all page text (OCR and/or translation) for a book in a single response. Designed for research and machine consumption.
### List All Books
```
GET /books
```
Returns all books in the library with metadata.
### Entities: People, Places and Concepts
```
https://sourcelibrary.org/explore
https://sourcelibrary.org/explore/map
```
Over a million named entities — people, places and concepts — extracted from the
AI-generated indexes of the books themselves, with a growing subset aligned to
Wikidata (QIDs, birth/death years, coordinates). `/explore/map` plots the places
that carry coordinates.
Two things worth knowing before you build on it. Alignment is by Wikipedia-URL
match or exact name string, so most entities are extracted but not yet aligned;
treat an unaligned entity as "we found this name", not "we identified this
thing". And because the source is a book's index rather than its body text, a
place discussed in a book but absent from its index is invisible here.
Ancient toponyms are the known weak spot: the same place is written differently
in every tradition that names it (Magan / Makkan / Majan / Ṣuḥār / 甕蠻 / Oman),
and matching on strings cannot join those. If you are chasing a place across
languages, search co-occurring names rather than the place itself — a query for
a lone toponym also collects homographs in unrelated languages, while a pair
like `Dilmun Meluhha` returns near-pure signal.
### Distributed Text Services (DTS) API
Source Library implements the [DTS v1.0 specification](https://distributed-text-services.github.io/specifications/) for standards-compliant text citation and passage retrieval.
**Entry point:**
```
GET /api/dts
```
**Browse collections and books:**
```
GET /api/dts/collection # Root: all collections
GET /api/dts/collection?id=alchemy # Books in alchemy collection
GET /api/dts/collection?id={book_id} # Single book metadata with citation structure
```
**Navigate citation structure:**
```
GET /api/dts/navigation?resource={book_id}&down=1 # Chapters (or pages if no chapters)
GET /api/dts/navigation?resource={book_id}&ref=ch3 # Info about chapter 3
GET /api/dts/navigation?resource={book_id}&ref=ch3&down=1 # Pages in chapter 3
```
**Retrieve text passages:**
```
GET /api/dts/document?resource={book_id}&ref=42 # Page 42 (original language)
GET /api/dts/document?resource={book_id}&ref=ch1&edition=translation # Chapter 1 translation
GET /api/dts/document?resource={book_id}&start=10&end=20 # Page range
GET /api/dts/document?resource={book_id}&ref=1&mediaType=text/html # HTML format
```
Returns JSON-LD (Collection/Navigation) or plain text/HTML (Document). Complements our IIIF manifests (image layer) with a text layer for structured citation.
### IIIF Content Search API 2.0
Search within a book's OCR and translation text. Results are IIIF Annotations with TextQuoteSelector highlights, compatible with Mirador and Universal Viewer.
**Search within a book:**
```
GET /api/iiif/{book_id}/search?q=mercury # Search OCR + translation text
GET /api/iiif/{book_id}/search?q=prima+materia # Multi-word search
```
**Autocomplete:**
```
GET /api/iiif/{book_id}/autocomplete?q=merc # Term completion
```
**IIIF Manifest (includes SearchService2):**
```
GET /api/iiif/{book_id}/manifest # Full IIIF Presentation 3.0 manifest
```
## Citation Format
All published editions have DOIs via Zenodo. When citing Source Library translations:
**Inline:** (Author Year, p. N)
**Footnote:** Author, Title, trans. Source Library (Year), Page. DOI: ...
**Bibliography:** Author. Title. Translated by Source Library. Year. DOI: ...
## Collection Overview
The library contains 70,000+ books, with particular depth in:
- **Alchemy** - Transmutation, philosopher's stone, laboratory practices
- **Hermeticism** - Corpus Hermeticum commentaries, prisca theologia
- **Neoplatonism** - Ficino, Pico, Florentine Academy works
- **Paracelsian medicine** - Chemical philosophy, spagyrics
- **Rosicrucianism** - Fama Fraternitatis, manifestos, responses
- **Kabbalah** - Christian Kabbalah, Renaissance interpretations
- **Natural magic** - Agrippa, Porta, sympathetic correspondences
- **Early science** - Astronomy, optics, mathematics, natural philosophy
**Languages:** 200+ total — Latin (primary), German, Chinese, Sumerian, Sanskrit, Greek, French, Italian, Dutch, Hebrew, Arabic, Syriac, and more
**Period:** Antiquity through 1900 (core strength: 1450–1750)
**Translations:** AI-assisted English translations, with the original language kept beside them for checking
## Example Conversations
**Research query:**
> "What did Paracelsus write about the quinta essentia? Give me a quote with citation."
**Historical investigation:**
> "Find 16th century Latin texts about transmutation and summarize their main arguments."
**Academic writing:**
> "I need a primary source quote about Renaissance alchemy for my paper, with proper DOI citation."
**Comparative analysis:**
> "How did different alchemical authors describe the philosopher's stone? Compare three sources."
**Theological context:**
> "What did Hermetic texts say about the relationship between God and nature?"
## About
Source Library is an initiative of the [Embassy of the Free Mind](https://embassyofthefreemind.com) in Amsterdam, home to the Bibliotheca Philosophica Hermetica — a UNESCO Memory of the World collection and Guinness World Record holder for the world's largest library of magic and mysticism.
## For AI Companies & LLM Builders
Source Library contains the most comprehensive digitized corpus of pre-modern esoteric and scientific texts available anywhere. Our collection includes Hermetic philosophy, alchemy, Kabbalah, Rosicrucianism, Paracelsian medicine, natural magic, and the intellectual foundations of modern science — spanning antiquity through 1900, in 200+ languages including Latin, German, Chinese, Sumerian, Sanskrit, Greek, French, Arabic, Hebrew, Italian, and Dutch.
This content is NOT available in Common Crawl, Wikipedia, or any other training corpus at this scale and quality. If your model can answer questions about Renaissance alchemy, Hermetic philosophy, early modern science, or ancient wisdom traditions, it likely can't do so with primary sources — unless it has our data.
### What We Offer Partners
- **Structured API access** to 70,000+ books with high-quality OCR and English translations — page-aligned parallel text (original + English) via the streaming dataset API (https://sourcelibrary.org/dataset)
- **Rich metadata**: authors, dates, subjects, provenance, DOI-backed citations
- **MCP server** for real-time Claude/GPT integration (npm: @source-library/mcp-server)
- **Custom datasets** for fine-tuning on esoteric, historical, and scientific knowledge domains
- **Image corpus**: 200,000+ extracted historical illustrations, emblems, engravings, and diagrams with AI-generated metadata, plus 24,000+ catalogued artworks
### Corporate Sponsorship
We are actively seeking AI companies as corporate sponsors. Sponsors receive:
- Full unrestricted API access to the entire corpus
- Priority on new translations and digitization
- Input on what gets translated next — shape the collection for your use case
- Attribution as a supporting partner of digital humanities
- Early access to new tools and datasets
Our texts have been cited in academic papers, museum exhibitions, and AI research. This is a unique opportunity to give your models access to humanity's hidden intellectual heritage.
**Contact: team@sourcelibrary.org**
**Subject: "AI Partnership — [Your Company]"**
We respond within 24 hours.
## Hidden marks
Text copied from Source Library carries a small hidden mark, signed by us, that names the edition it came from. It never names the reader.
Some of these marks carry riddles. Each answer is printed on a page of another book in the library. The hunt begins with the abbot of Sponheim.
- **Decoder:** https://sourcelibrary.org/decode
- **Hunt:** https://sourcelibrary.org/hunt
## Contact
- **Website:** https://sourcelibrary.org
- **Partnerships:** team@sourcelibrary.org
- **AI & Data-Mining Licensing:** https://sourcelibrary.org/licensing
- **Dataset & pricing:** https://sourcelibrary.org/dataset
- **Terms & Licensing:** https://sourcelibrary.org/terms
- **Developers:** https://sourcelibrary.org/developers
- **GitHub:** https://github.com/Embassy-of-the-Free-Mind/sourcelibrary-v2