AI Brand Visibility Report

Track how often your brand is mentioned and cited across ChatGPT, Claude, Gemini, and Perplexity when buyers ask AI assistants about your category. Get a 0–100 visibility score, share-of-voice vs. competitors, per-engine breakdown, an opportunity-gap list, and a client-ready HTML report — not just raw rows.

Keywords for store SEO (edit the store title/description to include these): AI visibility, GEO score, generative engine optimization, AEO, brand monitoring ChatGPT, AI search visibility, share of voice, Perplexity citations.

What it does

1. Generates realistic, brand-agnostic buyer prompts for your category (or uses your own).
2. Runs each prompt across the AI engines you enable.
3. Detects, per answer: whether your brand is mentioned, its position, whether your domain is cited, and which competitors appear.
4. Aggregates into a visibility score, share of voice, citation rate, per-engine and per-competitor breakdowns, and a gap list (buyer questions where competitors show up and you don't).
5. Outputs: a structured dataset (one row per prompt×engine), a REPORT HTML deliverable, and an OUTPUT summary object.

Inputs

• brand (required), brandDomain, brandAliases
• competitors (up to ~8)
• industry — used to auto-generate prompts, or supply customPrompts
• promptCount (default 25)
• engines — any of openai, anthropic, perplexity, gemini
• analyzeSentiment — optional extra pass
• API keys per provider (or set as environment variables / actor secrets)

Outputs

• Dataset — per prompt×engine: mentioned, position, cited, competitors seen, sentiment.
• Key-value REPORT — a self-contained HTML report (prints to PDF).
• Key-value OUTPUT — headline metrics as JSON.

Cost model (READ BEFORE PRICING)

Your cost of goods is LLM API calls: every prompt hits each enabled engine once (plus one cheap call to generate prompts, plus one per mention if sentiment is on). Estimate real token cost at your target promptCount × engines, then price the per-event rate so tokens stay well under ~⅓ of what you charge. Benchmark on the store is ~$40–70 / 1,000 prompts analyzed for this deliverable class.

Honest measurement caveat

API responses are not identical to the consumer apps (e.g. the OpenAI API is not the same as ChatGPT.com with live search). Perplexity's API is closest to real AI search because it returns citations. Scores are directional and best read as a trend over time — say this plainly in the listing; every serious tool in this space has the same caveat.

Run locally

npm install
export OPENAI_API_KEY=...        # + ANTHROPIC_API_KEY, GOOGLE_API_KEY, PERPLEXITY_API_KEY
apify run                        # or: npm start  (with an INPUT.json / apify CLI)

Monetization

Set chargePerPrompt: true only on the published, pay-per-event actor. The event name is prompt-analyzed (see src/config.js); configure the price in the Apify monetization settings to match.

Verify before shipping

Model IDs in src/config.js change often — confirm each against current provider docs.