fieldnotes runs agents that never stop reading what people say about your product across x and the open web — then turns what matters into github issues, mapped to the code responsible.
people feel they can't trust a session to survive their own workday. the tone shifted from confusion to open frustration on day four.
third time today i've been logged out of acme mid-deploy. losing my mind
Sessions invalidate while actively using the dashboard. Repro ~2h after login.
Great tool but I re-authenticate more than I deploy. Two stars until this is fixed.
agents sweep x continuously, with web, forums, reviews, and support threads flowing through the same source-agnostic pipeline.
hundreds of differently worded posts collapse into a handful of things people actually mean, every quote kept as evidence.
each theme carries a fourteen-day curve, so you can see patience wearing thin — and confirm when a fix actually landed.
volume, reach, recency, and confidence separate a stray grumble from an emerging product problem.
grok reads hundreds of source files in one pass and points each theme at the files and lines most likely responsible.
one click opens an issue with the summary, the user's own words, mapped code locations, and a suggested first investigation.
grok reads posts minutes after they're written, where opinions actually form. no scraping layer to maintain.
scanning runs for weeks, not seconds, so fieldnotes sees how feeling changes rather than one frame of it.
hundreds of source files read together means real file and line references, and no vector database to keep in sync.