Everything worth knowing about Scholaris
What Scholaris reads, how it finds and cites, what format it stores, what it does with your data, what it costs and how an agent uses it. More detailed than the front page, and written to be read slowly.
Reviewed on This page as Markdown
What this part is for
The front page says why Scholaris exists. These pages say how it works, with the specifics: which formats go in, where each page number comes from, which provider reads what, how long the things we measured take and what they cost, and what it does not do yet. If anything here disagrees with what you see in the app, the app is right and this page is wrong: write to us and we will fix it.
Every page has a Markdown twin (the same address ending in .md). There is an index for machines at /llms.txt and the full text of every page at /llms-full.txt. If you are an agent, start with the page for agents.
The pages
- What Scholaris is, and what it will never doA library that reads your sources and, when you ask, points to the exact printed page or second. The rules that hold it up: never invent a citation, show the provenance, leave the thinking to whoever writes.
- Which formats go in, and how each citation is anchoredDigital and scanned PDFs, photos, EPUB, Word, slides, spreadsheets, audio, video, web pages, YouTube and podcasts. For each, how it is read and which anchor it leaves: printed folio, second, slide, rows or paragraph.
- The SPDF 4.1 format, a library you can take with youAn .spdf is a gzip-compressed SQLite database holding the document already read: text, anchors, sections, figures, vectors and provenance. It opens without Scholaris and offline.
- How it searches, by words, by meaning and by imageThree searches at once (lexical, semantic and visual), fused and reranked; exact phrases in quotes; a modernised-spelling layer for Old Spanish and Latin; and results that arrive in two stages.
- Citations that are checked before they reach youNine CSL styles, BibTeX, RIS and CSL-JSON; an autocite that reads your draft and proposes each reference with its page; claim verification with temporal logic; and insertion into .docx without breaking the formatting.
- The tools for thinking with a whole libraryThe graph of people, works, places and concepts linked to Wikidata; the concept map; the citation graph between your books; watchers that tell you when something changes; and notebooks whose citations are checked again.
- Libraries that fill up at once and can be sharedFill a library with folders, ZIPs, lists of links or a Zotero BibTeX; export it as a .scholaris package; share it, follow it, copy it or publish it with a link, respecting its rights.
- Interviews, lectures and videos you can cite to the secondWord-by-word transcription with who is speaking, citable stretches of 30 to 60 seconds, searchable video frames, and a player with a synchronised transcript where you cite by selecting the text.
- API v1 and the MCP serverNine verbs over HTTP with one key, answers in JSON or Markdown, and an MCP server with OAuth for Claude, Cursor and other agents. Limits, scopes, errors and examples to copy.
- Scholaris on your computer, or in your own Cloudflare accountThe same app with SQLite and your disk: on Node, in Docker, as a desktop executable or deployed to your Cloudflare account. What stays on your machine and what still needs the cloud.
- What happens to your data, provider by providerWhere your documents are stored, which AI provider reads what, what is sent to the open bibliographic databases, how your keys are encrypted and what is deleted when you delete.
- Plans, coupons and limitsA free plan to start, Pro for real work, and the home version with no quotas. How much fits in each, how pages and minutes are counted, and how coupons work.
- What we measured, and under what conditionsReading times and costs, folio accuracy, search quality, invented citations and latency, as they came out of the benchmark of 6 October 2026, and what those figures do not say.
- Frequently asked questionsWhether it invents citations, whether it works with seventeenth-century scans or video interviews, what happens to your data, whether it works offline, and other questions, answered plainly.
- Scholaris glossaryThe words Scholaris uses, from folio and anchor to pliego, imprenta, SPDF or watcher, with what they mean here and, where it helps, in the tradition of the book.
- Scholaris beside Zotero, Elicit, NotebookLM and the restWhat each tool does, when another one suits you better than Scholaris, and what Scholaris does that the others do not set out to do. No figures about others that we cannot check.
- ChangelogWhat has changed in Scholaris, from the first version to the second, told by milestones and in the order they were made.
- For agents: how to read Scholaris and how to use it on someone’s behalfWhat an agent can read on this site and in which format, how to act on a person’s library through API v1 or MCP, with which permissions and limits, and the rules for citing without inventing.
Scholaris in three sentences
- It reads what you give it (digital and scanned PDFs, photos of pages, EPUB, Word, slides, spreadsheets, audio, video, web pages and YouTube) and stores an anchor for every passage: the printed page, the second, the slide or the paragraph.
- When you search or ask, it gives you passages with a citation ready to paste, and every citation is written from the stored anchor, never from what a model says.
- It does not write for you or decide which author is right: it points to the place and steps aside.
Who makes it
Scholaris is made by José Luis Saorín Ferrer, philologist and programmer, in Santa Cruz de Tenerife. Write to jl@joseluissaorin.com.