GEO Radar
Customers ask AI. Are you in the answer?
A four-step loop: monitor, diagnose, optimise, review. The prompt library separates brand, product, scenario and competitor intents and expands in batch; every prompt is run engine by engine and the answer text, mention status and cited domains are stored. Coverage is 18 engines (13 domestic, 5 overseas) across API, browser and simulated modes, with same-prompt competitor comparison and citation-domain ranking.
On the optimisation side it drafts GEO articles, FAQs, landing pages and knowledge feeds; a site health check covers robots for nine AI crawlers plus structured-data validation, with a distribution matrix and effect tracking; changes are re-measured with the same prompts. The alert centre pushes falling mention rates, brand disappearance, competitors overtaking and negative spikes into in-app notices and group robots; the report centre issues weekly and monthly reports, HTML exports and expiring share links. Real versus simulated results are always labelled honestly - without a key the system falls back to a deterministic simulator instead of faking answers. An open API offers workspace keys with read-only or read-write scopes, and private models can be connected.
It is live on the public internet, with desktop browser automation keeping publishing and collection on the client's own machine.
Core capabilities
- 18 engines (13 domestic, 5 overseas) in API, browser and simulated modes
- Intent-tagged prompt library with batch expansion
- Stores answer text, mentions and cited domains
- Same-prompt competitor comparison and source-domain ranking
- Site health check for nine AI crawlers and structured data
- Distribution matrix with re-measurement
- Alerts to in-app notices and group robots
- Weekly and monthly reports, HTML export, expiring share links
- Open API with workspace keys and read-only or read-write scopes
- Honest real-versus-simulated labelling, never faked
Who uses it
- Brand marketing teams tracking mentions and negatives in AI answers
- Agencies running many clients from one console
- Visibility of global brands on overseas engines
- Before-and-after measurement for a product launch
Delivery models
Hosted SaaS · on-premises · private model integration · source licence
Tech stack
Node.js 20+ + Fastify + BullMQ · React + antd + ECharts · uni-app · Electron · MySQL / MongoDB / Redis / RabbitMQ