Pain map
What breaks in document AI
290 production incidents from public threads. Every quote links to its source. Pick a category and check it yourself.
Lost edges and provenance
5 of 5 quotes
“No `EdgeType cleanup failed` warning is logged; the drift comes from the cleanup logic, not from an exception.”
GitHub, cognee · 7 October 2026 · open the thread “So an edge that is not the first of any node gets dropped, whatever the token budget.”
GitHub, graphrag · 30 September 2026 · open the thread “I collected 83 dead-lettered episodes carrying this error from a production deployment”
GitHub, graphiti · 12 September 2026 · in production · open the thread “For a memory server this is silent data loss: the client believes the fact was stored, and nothing in the response, the tool result, or any queryable state says otherwise.”
GitHub, graphiti · 30 July 2026 · in production · open the thread “This blocks any production use of Cognee with non-trivial data volumes on the default SQLite backend.”
GitHub, cognee · 24 April 2026 · in production · open the thread Hit this one too? We are building for it. Join the waitlist.
Method
We collected 448 public posts published after 1 November 2025 in which a builder describes document AI failing. Each link was re-opened on 8 November 2026. We kept the 290 where the quote is on the page, the author describes an incident and not an opinion, and has no tie to a vendor. Quotes are verbatim and trimmed after 420 characters. The grouping is ours. A complaint is not a measurement.