The open-source apps featkpr has run on, and how far each got.
1 through all six steps · 1 through two · 7 lined up · as of 27 Sep 2026
BookStackopen-source wiki 307 features in 12 modules · 163 paths to 69 goals · 304 tests pass, 269 refused to the wrong user · 605 of 2,372 missed in the missed-items check
Shlinkself-hosted URL shortener 28 routes read · 20 features named · the crawl and tests need a reader for its stack, which comes next
featkpr itselfthis product, on its own code dogfooding stage 1 built: 26 of its own screens entered as planned rows · reading its own merge requests, crawling and testing itself are planned
Uptime Kumauptime monitor the TypeScript indexer read 8,475 symbols in 388 files in 10 s · 0 of 113 routes keyed yet · first target for the Node route reader
Paperless-ngxdocument manager the Python indexer read 405 of 405 files, 18,867 symbols · routes need the Python route reader (planned)
InvenTreeinventory management a chosen next-stage target, with its own package and pass criteria
Immichphoto library a chosen next-stage target, with its own package and pass criteria
Formbrickssurveys a chosen next-stage target, with its own package and pass criteria
Actual Budgetpersonal finance picked as the product a non-technical person already has opinions about · nothing run yet
16 more weighed in research, not started
Chatwoot
kept as an alternate: named feature flagscustomer support
Documenso
kept as an alternate: 132 end-to-end specs to compare againstdocument signing
OpenProject
on the candidate listproject management
Mastodon
on the candidate listsocial network
Weblate
on the candidate listtranslation
Discourse
on the candidate listforum
Taiga
on the candidate listagile boards
Outline
on the candidate listteam wiki
Cal.diy
dropped: maintenancescheduling
Medusa
dropped: six times the sizecommerce
Baserow
dropped: hosted on GitLab, split tiersno-code database
Plane
dropped: no end-to-end suite to compare withproject tracking
Twenty
dropped: per-object GraphQL APICRM
Zulip
dropped: few screensteam chat
Sylius
dropped: no end-to-end suitecommerce
Firefly III
dropped: no end-to-end suitepersonal finance
the reasons are the research's own · our research on which products to map first
How a product joins
The same route for each one, and the audit says what was missed before anything is published.
1 · a folderDescribe the targethow to check it out, start it and seed one user per role
2 · a readerRead its stackLaravel and Mezzio are read fully today; Node and Python readers are planned
3 · one paid stepName the featuresthe naming pass, about a dollar for a new product
4 · a copyCrawl, draft, teston a throwaway copy, never the real product
5 · the auditCount what's missingthen publish, with the misses printed
Want your product in the library? Name it in the demo form →