Product status

What is collecting real data today, what is still a preview, and what is next.

Geonimo is being built in the open. This page is the honest map of where the platform stands, so nothing in the rest of these docs reads as a promise it cannot keep.

Collecting and reporting real data

The collection engine is live. Prompts are scraped daily against real consumer AI surfaces, answers are ingested, mentions are extracted and metrics roll up on ingest.

ScreenState
OverviewReal — KPIs, competitor table, visibility trend, recent answers
SourcesReal — domains and URLs, quoted vs listed
Query Fan-outReal — the searches engines actually ran
Ads & ShoppingReal — carousel and ad placements
Brand PerceptionReal — awareness, quotes, co-mentions, sources, and strengths / weaknesses / misconceptions with verbatim evidence
PromptsReal — generation, CRUD, per-prompt metrics
CompetitorsReal — CRUD and mention counts
ProductsReal — product-level visibility
PR OpportunitiesReal — deterministic gap detector

Accounts, teams, roles, invitations, billing, authentication and MFA are live, as is the marketing site, blog and these docs.

Preview screens

These are built and navigable but not yet wired to your data. They render a fixed sample dataset for a demonstration brand:

ScreenWaiting on
RecommendationsThe recommendation generator over the gap detectors
Pages OptimizationThe page audit crawler
GEO CopilotThe tool-use agent over your workspace data
PagesThe page audit crawler
AI TrackerThe tracker ingest endpoint

Telling them apart is easy: a preview screen shows the demonstration brand set (Geonimo, Profound, Peec AI, Otterly…) rather than the brand and competitors you configured.

Partially wired

A few elements inside live screens are still sample data, and are captioned as such in the interface:

  • Citation Index on the Overview — citations are not yet attributed to individual entities.
  • Platform, location and tag filters on the Overview — the filters render but do not yet scope real data.
  • Sentiment and recommendation strength per mention — not yet scored, so sentiment figures come from the sample layer. Perception claims — strengths, weaknesses and misconceptions with verbatim evidence — are live and separate from this.

What runs behind the screens

  • Collection — daily dispatch at 01:00 UTC plus an hourly catch-up, submitted to the scraper in batches, ingested by callback, with a reconciliation sweep every five minutes. See Analysis pipeline.
  • Engines — three scraped consumer surfaces collected daily (ChatGPT, Google AI Overviews, Perplexity), plus Google AI Mode. See AI engines.
  • Measurement — mentions extracted deterministically from answer text, rolled up on ingest into daily metrics with Wilson confidence intervals. See Metrics.
  • Isolation — every GEO table is behind row-level security, with tenant-isolation tests in the migration suite. See Security and privacy.

Next

  1. Per-mention sentiment and recommendation-strength scoring.
  2. Entity-attributed citations, retiring the sampled Citation Index.
  3. Engine, location and tag filters scoping live data across every tab.
  4. Page audits, feeding Pages and Pages Optimization.
  5. Recommendations generated from the deterministic gap detectors.
  6. Tracker ingest, for AI referral traffic.