Questions, answered

It groups by meaning, not keywords. Every entry becomes an embedding, and recurring complaints cluster into emergent themes — so problems like Checkout or Search surface on their own — ranked into a Top 5, each with real sub-issue counts and customer quotes.

No — that's the point. Upload a CSV or paste text, and GliddeSignal normalizes entries, drops empty rows, removes duplicates, and embeds everything on upload. For a messy CSV it even auto-detects the column that holds the feedback.

No. An intent gate classifies every entry as a complaint, praise, suggestion, or neutral before anything is grouped. Only genuine problems feed the themes, so counts stay honest.

A ranked Top 5 of problems — each with quotes and sub-issue counts — a feature-level pain map, a view of what changed since your last run, a Decision Assistant with a sprint-sized action per problem, and a one-page executive memo you can export as a branded PDF.

Whenever you have a fresh batch — many teams run monthly. The 'What Changed' view only appears once there's a previous run to compare against, so a regular cadence is where GliddeSignal becomes most useful.

You correct outputs inline — re-tag a quote, rename a problem, edit or 👍/👎 an action or the memo. Each correction is remembered for that project and quietly steers future runs toward your vocabulary. There's no model fine-tuning, and with zero corrections behavior is unchanged.

Every top problem comes with actions you can copy straight into Jira, Linear, or Slack, plus a downloadable one-page memo PDF you can share with leadership.

Your corrections stay scoped to your own project and there's no model fine-tuning on your data. Every request enforces per-account ownership on your projects, runs, and feedback.

Starter is $99/mo (3 projects, 3 runs/month, last 10 runs). Pro is $199/mo (5 projects, 10 runs/month, last 50 runs). Annual billing gives you two months free. Runs reset monthly on either interval.