Skip to main content
Stay updated on new features, improvements, and changes to the Urantia Papers API ecosystem.

November 2026

November 15, 2026

Spanish review flags

  • We read every Spanish paragraph that the automatic checks flagged. 30 paragraphs got fixes and 2 were written again, and no paragraph has translation.review: "needs_review" now. See Translations

November 14, 2026

Editions in MCP

  • New MCP tools: editions.list (the editions and their counts of changes) and editions.variants (the changes since 1955, by paper, paragraph, kind, or edition). The MCP server has 28 tools. See Editions of the text
  • GET /editions/variants takes edition: the changes adopted in that edition. from and to give each change between two editions, such as from=1955&to=srt-2015, and meta.applied counts the changes that the default text has
  • Verify quotes adds match.editions: each edition that has the quoted wording, oldest first
  • The MCP answer to initialize has instructions and a server description
  • A format change no longer has an empty <em></em> in before or after
  • @urantia/api 0.15.0: api.editions.variants takes edition, from, and to, and its meta has applied

November 11, 2026

Editions of the text

  • edition on GET /paragraphs/{ref}, /paragraphs/{ref}/context, /paragraphs/random, and /papers/{id} gives the text of an older edition: 1955 or a Standard Reference Text from 2009 to 2021. The default stays srt-2024. See Editions of the text
  • include=variants gives each change to a paragraph since 1955: the words before and after, the year, the 1955 page, and the reason
  • New endpoints: GET /editions and GET /editions/variants
  • Each answer of these routes, and of search, has meta.edition
  • The MCP tools paragraphs.get, paragraphs.context, and papers.get take edition and include_variants
  • @urantia/api 0.13.0 adds the edition option, include: 'variants', and api.editions
  • Verify quotes knows the editions: a quote of an older wording, such as the 1955 text, is exact in its edition, with match.edition and match.variants
  • @urantia/api 0.14.0: the Paper and PaperDetailResponse types match what the API answers (data.paper and data.paragraphs)

November 9, 2026

2024 text corrections and safe note changes

  • The text now has the 2 corrections that Urantia Foundation approved in 2024: 98:3.8 reads “two-score”, and 133:7.9 reads “associator”. The text is now the same as the Foundation download 1.22. See Edition of the text.
  • PUT /me/notes/{id} requires the updatedAt that your app read. If the note changed after that, the answer is a 409 with the note as it is now in current. @urantia/api 0.12.0 throws NoteConflictError. See Change a note that another app can change.
  • Two saves of a bookmark of the same paragraph, at the same moment, make one bookmark.

Text edition, scripture parallels in MCP, and kept deletes

  • Edition of the text: the API serves the Standard Reference Text, downloaded from Urantia Foundation in January 2024.
  • The MCP paragraph and search tools take include_scripture_parallels: the 3 passages nearest in meaning from each of the 21 world religions texts. @urantia/api 0.11.0 has the ScriptureParallel type.
  • Sync keeps at most the newest 1,000 deletes of each kind for a person. See Sync bookmarks and notes.

Read-only permissions and sync

  • New permissions bookmarks:read, notes:read, reading-progress:read, and preferences:read allow GET only. See Permissions.
  • GET /me/bookmarks and GET /me/notes take since, and answer syncedAt and deleted. See Sync bookmarks and notes. @urantia/api 0.10.0 has the option, and its list types now say pagination, which is what the API answers.
  • A person can download a copy of their account data on their account page.
  • The @urantia/auth page has a section on data, security, and terms, and the Privacy Policy says what we do after a breach.

What a parallel is

  • How the links are made now says what a parallel is and is not, how the API computes it, what the similarity number means, and how computed parallels differ from a list made by hand.

October 2026

October 29, 2026

How the Spanish text was made

  • Translations says why the text is CC0, how it was made, how a blind review scored it beside two published editions (accuracy 4.93 of 5, no major errors in 80 paragraphs), and what the limits are

AI Assistants names the other texts

  • The AI Assistants page says that an assistant can read the 21 other texts beside the Bible, with an example question for the Bhagavad Gita

October 26, 2026

Spanish text corrections

  • The names of peoples are lowercase, as Spanish writes them (“los noditas”, “los anditas”), in 264 paragraphs, 23 titles, and the entity names; “orden” is masculine for an order of beings in 144 paragraphs; and 19 other paragraphs have small fixes. If you keep a copy of the Spanish text, fetch it again

A page for each text

  • Each of the 21 world religions texts has its own page under “Each text”, such as the Quran, the Bhagavad Gita, and the Tao Te Ching: how to get a passage, how its references work, search, and its source
  • The Bible page starts with a quick start (a verse, a chapter, and the books with their chapter and verse counts)
  • The removed Concepts and Quotes pages now answer 404 here, not a redirect to the home page

October 24, 2026

The works catalog in Spanish

  • lang=es on /works, /works/{work}/{ref}, /scriptures, /scriptures/{corpus}, and /bible/books gives the titles, traditions, summaries, labels, chapter titles, and Bible book names in Spanish, with a catalogTranslation label. The source works are English, and their text stays English (textLanguage). See Translations
  • The MCP tools works.list, works.read, and bible.books take lang: "es"
  • @urantia/api 0.9.0 adds lang to works.list, works.read, and bible.books

October 23, 2026

  • /search and /search/semantic take lang=es, and so do the MCP tools search.fulltext and search.semantic. See Translations
  • With include, the Papers parallels and the entity names come in Spanish
  • Entities: 2,517 entities now have a Spanish description
  • @urantia/api 0.8.0 adds lang to search.fullText and search.semantic

Spanish

  • All 197 papers are in Spanish: a machine translation, released as CC0. Add lang=es to the papers, paragraphs, and table of contents endpoints. Each translated paragraph has language and a translation label. See Translations
  • The MCP paper and paragraph tools take lang: "es"
  • @urantia/api 0.7.0 adds lang to papers.list, papers.get, and toc.get, and the TranslationLabel type
  • Entities: Spanish now has entity names only (2,519), with English descriptions

October 21, 2026

  • The documentation site now counts page views with PostHog. The Privacy Policy says what is recorded

October 20, 2026

  • Urantia Papers is a ChatGPT plugin. AI Assistants now starts the ChatGPT steps with Install plugin; developer mode stays as the second way
  • AI Assistants: the ChatGPT steps add Set up Urantia Papers, the step after the install that loads the research skill and checks the connection

October 18, 2026

Works: one list for the Bible and every text

  • Added /works: the Bible and the other texts in one list and one result shape, with keyword and semantic search, and the MCP tools works.*. Each work names its tradition, its kind (scripture or classic), and a one-sentence summary
  • Added twelve public domain texts: the Book of Mormon (1830), the Rig Veda (Griffith, 1896), the Egyptian Book of the Dead (Renouf and Naville, 1904), Hesiod and the Homeric Hymns (Evelyn-White, 1914), the Poetic Edda (Bellows, 1923), Aztec hymns (Brinton, 1890), Chuang Tzu (Giles, 1889), the Prose Edda (Brodeur, 1916), the Kojiki (Chamberlain, 1882), three Upanishads (Muller, 1879), and the Shû King (Legge, 1879). See World religions
  • Traditions now use one form for every work: “Hindu”, “Buddhist”, “Taoist”, “Confucian”, “Muslim”, “Sikh”, “Cynic”, “Japanese (Shinto)”, and “Jewish and Christian” for the Bible

October 15, 2026

Audio for a whole paper

  • GET /papers/{id} now returns paper.audio: the MP3 file of the whole paper, and the start of each title, paragraph, and sentence in it. See Audio
  • New recordings for eight section titles and for ten “Presented by” and “Sponsored by” lines where words were missing, for example “Taxation” (72:7) and 78:8.13

October 7, 2026

Patterns across the texts

  • Added four lists under /scriptures/insights: shared currents, mutual pairs, leans, and far from the texts. See World religions
  • Added ?include=scriptureScores on paragraphs

World religions texts, phase 2

  • Added five public domain texts: Diogenes Laertius, Book 6 (Hicks, 1925), and Epictetus, Discourses 3.22 (Oldfather, 1928), for Cynicism; Shinto oracles quoted by Aston (1905); the Koran (Pickthall, 1930); and the Japji (Macauliffe, 1909). See World religions
  • References can now have three levels, such as Epictetus 3.22.45. Each text states its refLevels and has notes on how it is numbered
  • ?include=scriptureParallels now returns the 3 closest passages from each of the nine texts

October 6, 2026

World religions texts

  • Added four public domain texts of the religions in Paper 131: the Dhammapada, the Tao Te Ching, the Analects, and the Bhagavad Gita. See World religions
  • Each passage links to the closest Urantia paragraphs, and each paragraph links back with ?include=scriptureParallels
  • Added GET /scriptures, GET /scriptures/{corpus}, GET /scriptures/{corpus}/{ref}, GET /scriptures/{corpus}/{ref}/urantia-parallels, and POST /scriptures/search/semantic

Verify a quote

  • Added POST /quotes/verify. It checks whether a quoted passage is in the Urantia Papers, and returns the reference, the exact text, and each word that differs. See Verify a quote
  • The lang parameter now says that paragraph text is English only. Entity names and descriptions are translated

October 5, 2026

OpenAPI spec

  • Every operation in the OpenAPI spec now carries an example. The response examples are real output from the API
  • Every error response references one shared ProblemDetails schema (RFC 9457)
  • The spec documents the X-RateLimit-* headers on every response, and the 429 response
  • Added an API catalog at /.well-known/api-catalog (RFC 9727)
  • robots.txt on urantia.dev and api.urantia.dev now states that search, AI answers, and AI training are all allowed

October 4, 2026

AI assistants

  • Added AI Assistants: steps to connect Claude, ChatGPT, Grok, or Gemini to the Urantia Papers, with no code
  • The plugin now installs as a Gemini CLI extension. See MCP Servers
  • Rewrote MCP Servers with the install commands for Claude Code, Cursor, and Gemini CLI
  • Published a Postman workspace with 28 requests that run as is

October 3, 2026

Documentation

  • Removed the Concepts and Quotes sections and two blog posts. The docs now cover the API, the MCP server, the SDKs, and tutorials. Old links to those pages redirect to the docs home
  • Renamed the Blog tab to Tutorials
  • Added Help and Contact

October 2, 2026

Feedback endpoint

  • Added POST /feedback for bug reports, docs gaps, and ideas. No authentication required
  • Limit of 5 requests per minute and 10 per 15 minutes per IP address
  • See the Feedback guide and the API reference
  • @urantia/api 0.5.0 adds api.feedback.submit (SDK docs)

Documentation & Developer Experience

MCP Server Expansion
  • Expanded MCP server from 13 to 19 tools with full Bible integration
  • Added 6 new Bible endpoint tools:
    • bible.books — List all Bible books
    • bible.book — Get specific Bible book details
    • bible.chapter — Retrieve Bible chapter content
    • bible.verse — Get specific Bible verse
    • bible.verse.urantia_parallels — Find Urantia paragraphs for a Bible verse
    • bible.search.semantic — Semantic search across Bible content
  • MCP server now available via Smithery with one-click installation
Developer Onboarding Improvements
  • Added comprehensive Daily Quote Bot Tutorial with examples for:
    • Discord, Slack, email, and Twitter/X integrations
    • Multiple scheduling approaches (GitHub Actions, Cloudflare Workers, cron)
    • Enhancement patterns (topic filtering, audio narration, history tracking)
  • Enhanced Quickstart guide with multi-language SDK examples:
    • Added tabbed code samples showing cURL, TypeScript SDK, JavaScript, and Python
    • Integrated demo.urantia.dev references throughout
    • Added “Next Steps” section for developer onboarding flow
Documentation Clarity
  • Clarified alias branding: urantiapapers.com and openurantia.com redirect to UrantiaHub
  • Updated ecosystem documentation to reflect correct platform relationships
  • Improved OpenUrantia.com repository descriptions

Initial Release - September 2026

Core API Features

REST API Endpoints (19 total)
  • Full access to 197 papers, 1,626 sections, 14,500+ paragraphs
  • Full-text and semantic search capabilities
  • Bible integration with cross-reference support
  • Audio narration endpoints with multi-voice support
  • Citation generation and OG image creation
  • Embeddings export for AI applications
SDKs & Developer Tools MCP Server (Initial 13 tools)
  • Core content access (papers, paragraphs, entities)
  • Search capabilities (full-text and semantic)
  • Audio narration integration
  • Citation and reference lookup
  • Compatible with Claude, Cursor, and other AI agents
Documentation
  • Complete API reference with interactive examples
  • Comprehensive guides for papers, paragraphs, entities, audio, and Bible data
  • Use case examples and integration guides

Note: This changelog tracks documentation, API, and tooling updates. For underlying platform updates, see the UrantiaHub repository.

Future Updates

Looking for something specific? Check our: