Feature ideas backlog

Ranked by gut-feel value:effort. Not committed to any of these.

Top 5 — high value, manageable effort

  1. Inline graph_data in bundle URL — already shipping. The decision write-up is at demo-dec-bundle-graph-inline. This is the load-bearing feature for the cross-AI thesis.
  2. Custom hub theme — let users pick an accent colour and a typeface. Modest engineering effort, real perceived-quality bump. Ship-able in a week.
  3. Slack capture — send a Slack DM to a bot, get a permanent mdfy URL back. Reuses the existing capture pipeline. Two weeks of work, mostly OAuth dance.
  4. Email-in capture — forward any email to a magic address, get a mdfy URL back. Cheaper than Slack because there's no OAuth, but spam-protection matters.
  5. Hub-to-PDF export — generate a single PDF of all public docs in a hub. Useful for portfolio-mode users. Most of the work is print CSS, which we already have for individual docs.

Mid-tier — interesting but later

  • Concept-page editing. Right now the concept index pages are auto-derived. Letting the user pencil-edit a concept's description is a big-feel feature with light engineering cost.
  • Per-bundle layouts. Bundle owners pick how the canvas paints (timeline / tree / grid). Adds dimensionality to the bundle as a surface.
  • Embed a doc inside another doc. The {{include: doc-id}} macro. Useful for compositional notes. Engineering risk is recursion.
  • Activity dashboard. Per-hub views/fetches/captures. Currently planned as a Pro feature.

Bottom of mind — fun, not strategic

  • Mobile native app.
  • Apple Watch app.
  • Voice capture with Whisper.
  • AR document spatial layout. (Joke. Mostly.)
  • AI-generated cover images for hubs.
  • Custom CSS per hub.

What I'd kill from the backlog

  • Real-time collab. The cost is enormous; the demand signal is weak.
  • Tasks / kanban. Different product.
  • Notion-block-style embedded widgets. Different product.

Decision rule

Anything in "top 5" only ships if the feedback from launch tells me it's the right next thing. The plan is a hypothesis; users will tell me which item to lift to actual roadmap.