Skip to content

Roadmap

This roadmap is a living document. Items are not promises; they reflect the maintainers' current intent.

The strategy is subscription-first: anyone with a normal AI subscription should get full value from this site without API keys, CLIs, or code. Developer Mode is a real path, but it is a path users opt into.

Done in v0.5 (subscription-first expansion)

  • Beginner / Power User / Builder / Developer mode selector in the Task Builder.
  • Provider-specific packages ("Use this in Claude / ChatGPT / Gemini / Grok / Perplexity / Copilot / Other").
  • Practice exercise + "what good output looks like" + Plans + Maximize-my-subscription + Expert-expansion outputs.
  • New top-level sections: Capability map, No-code automations, Memory & preferences, Examples, Learning path.
  • 13 beginner examples + the universal 7-mission learning path.
  • Plain-English glossary with schematic diagrams for ChatGPT Project / Task, Claude Project / Skill / Cowork, Gemini Gem / Scheduled action, Grok chat.
  • Traceability files updated; all 16 user requirements + the 24 acceptance criteria mapped.

Now (next ~30 days)

  • Replace schematic SVG diagrams with annotated screenshots on a quarterly cadence (auth-redacted).
  • Add 5 more beginner examples driven by reader requests.
  • Verify every product page against current vendor docs and refresh the Last verified: line.
  • Add a per-page Plan-availability shorthand legend at the top of every Mastery page (consistency pass).

Next (~30–90 days)

  • Localization: a parallel French + Spanish track for the homepage, learning path, and capability map.
  • Add a chapter on observability for subscription users — what each app surfaces about a chat (logs, share links, audit trails), without API access.
  • Expand the No-code automation guide with platform-specific cost-cap guidance (token caps, message caps, runs-per-day).
  • Add an "AI for non-knowledge-work" examples track (parents, drivers, tradespeople, students at any age).

Later

  • A self-graded eval set per provider that users can run by hand to score their AI on their own work.
  • An offline / disconnected mode page for users in low-bandwidth or air-gapped contexts.
  • A self-hosted alternative to GitHub Pages for organisations that cannot publish externally.

Out of scope (for now)

  • Vendor-specific deep dives that duplicate official docs. We summarise and link.
  • Reviews of unreleased products. We wait for GA + a public doc page.
  • Anything that requires running code on a schedule by default. Schedulers stay opt-in and manual; native scheduled actions are documented but never recommended without a 3-run drill.
  • Recommending API keys or developer tooling to users who said they're beginners.

Contributing to the roadmap

Open a GitHub Issue with the label roadmap and a short proposal. Material changes land as Architecture Decision Records under decision-records/.