AFFiNE
A nightly canary train where the AI platform work hides between dependency bumps
A side-by-side editorial comparison of Document360 and tufte — release velocity, themes, recent moves, and the top alternatives to consider.
| Feature | Document360 | tufte |
|---|---|---|
| Sector | Collab | Collab |
| Velocity score | 6.3 | 0.0 |
| Sparks · 30d | 1 | 0 |
| Top themes | api, oauth, mcp, knowledge base | r-markdown, typesetting, bookdown, pandoc-compat |
| Last editorial update | 3d ago | 1h ago |
| Website | — | Visit → |
Document360 rebuilds its public API and turns docs into an agent-readable surface.
Document360 ships a dated release every two to four weeks and the last six months have been dominated by two threads: making the knowledge base machine-consumable, and hardening enterprise access control. The AI thread runs from the March MCP server through publishing and workflow tools in June and per-reader-group AI restrictions this month; the access thread runs SCIM provisioning, multiple JWT configurations, and reader permission inheritance. This release adds a ground-up API v3 with OAuth 2.0 and scoped keys.
After three quiet years, tufte 0.15 finally lets Tufte-style documents cross-reference.
Almost every release before the latest is a Pandoc compatibility fix — sidenotes breaking on Pandoc 2.15, margin references misplaced on 2.11, table and citation handling in `tufte_handout()`, `--wrap preserve` for margin content on 2.19. Then 0.15.0 in May 2026 breaks the pattern: `tufte_handout2()`, `tufte_book2()` and `tufte_html2()` wrap the corresponding bookdown formats, enabling text references and cross-references in captions, and `margin_fig_pos` gives global control over margin-figure placement.
Document360 ships a dated release every two to four weeks and the last six months have been dominated by two threads: making the knowledge base machine-consumable, and hardening enterprise access control. The AI thread runs from the March MCP server through publishing and workflow tools in June and per-reader-group AI restrictions this month; the access thread runs SCIM provisioning, multiple JWT configurations, and reader permission inheritance. This release adds a ground-up API v3 with OAuth 2.0 and scoped keys.
The product is being rebuilt around programmatic access rather than portal usage. MCP made the knowledge base writable by an assistant, llms.txt made it discoverable to crawlers, copy-as-Markdown and open-in-ChatGPT made articles portable, and API v3 now gives all of that a permission-aware substrate the old v1/v2 endpoints could not support. The same permission model is showing up on the reader side too — Eddy AI features are now gated per reader group — so authorization is becoming the shared spine across API, AI, and reader access.
Expect the advanced API v3 endpoints to keep expanding as a paid add-on and the MCP server to be re-plumbed onto v3's scoped-key model, since MCP currently sits outside the new authorization scheme. A v1/v2 deprecation notice is the other likely follow-up.
Almost every release before the latest is a Pandoc compatibility fix — sidenotes breaking on Pandoc 2.15, margin references misplaced on 2.11, table and citation handling in `tufte_handout()`, `--wrap preserve` for margin content on 2.19. Then 0.15.0 in May 2026 breaks the pattern: `tufte_handout2()`, `tufte_book2()` and `tufte_html2()` wrap the corresponding bookdown formats, enabling text references and cross-references in captions, and `margin_fig_pos` gives global control over margin-figure placement.
The package spent years absorbing upstream Pandoc churn without adding anything, and has now moved to sit on top of bookdown rather than reimplement around it. That is the sustainable direction for a format package this small: inherit cross-referencing and text references instead of maintaining a parallel implementation. Release intervals remain long, so this is unlikely to become a fast-moving package.
Expect the bookdown-backed formats to absorb further fixes as the recommended path, with the plain `tufte_*()` formats staying frozen against Pandoc changes.
Other Collab products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either Document360 or tufte.
A nightly canary train where the AI platform work hides between dependency bumps
pagedown is down to one small fix a year, mostly absorbing changes from elsewhere.
xaringan has settled into theme upkeep and Pandoc-era compatibility patches.
Double is turning its accounting assistant into the way the work gets done
litedown grows from a lean renderer into a book and site generator
bookdown's output is mature; its free hosting destination is being switched off.
See all Document360 alternatives → · See all tufte alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. Document360 is currently shipping more aggressively (velocity 6.3 vs 0.0), with 1 editorial sparks in the last 30 days against 0. See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. Document360 is currently shipping more aggressively (velocity 6.3 vs 0.0), with 1 editorial sparks in the last 30 days against 0. For your specific use case, the alternatives sections above list other Collab products to evaluate alongside.
Top Document360 alternatives in Collab are ranked by recent ship velocity. Browse the "Document360 alternatives" section above for the current picks, or visit /alternatives/document360 for the full list with editorial commentary on each.
Top tufte alternatives in Collab are ranked by recent ship velocity. Browse the "tufte alternatives" section above for the current picks, or visit /alternatives/tufte-r for the full list with editorial commentary on each.