Reading log
Summary
Clusters
§ Brief
Fairmark started as a complaint about reading-list debt. We had thousands of saved articles and no system for re-encountering them.
Most read-it-later apps pretend to summarize. They confidently flatten a 9,000-word essay into three bullets and you never know what was lost. Fairmark takes the opposite approach: it tells you exactly which sections it summarized, which it skipped, and where it's guessing.
§ AI use
- Anthropic Claude handles section-level summarization with explicit coverage tags.
- Embeddings cluster reading logs; the user approves or rejects clusters.
§ Results
- 1,200+ active readers in private beta
- Average summary coverage: 73%
§ Limitations
- Doesn't yet handle PDFs cleanly
- No mobile app