Key takeaways
- Visby is an AI visibility tracker for ChatGPT, Claude, Gemini and Google AI Overviews. It starts at $79/month with a free trial, and review scores are very high (4.8/5 on G2 from 109 reviews, 4.8 on Trustpilot).
- Setup is the praised part: three onboarding steps, Google Analytics, Search Console and Bing connections, and a prioritized task list.
- The limits are real. You pick only 3 of 4 engines on every plan, Perplexity is not listed, and the answer quotas work out to roughly one run per prompt per engine per month.
- Reviewers report manual cleanup of competitor lists, an empty Reviews area, and no Traditional Chinese support.
- Good for a startup or solo marketer who wants a first look at AI visibility. Teams that need crawler logs, deep citation detail or automated publishing should compare other options.
What Visby AI is
Visby describes itself as an AI visibility intelligence platform. It shows where your brand appears, or doesn't, in ChatGPT, Claude, Gemini and Google AI Overviews. It then turns the findings into a list of GEO tasks and can generate articles to fill gaps.
Our catalog lists it as basic AI mention tracking for startups, and that's a fair summary. The pitch is simple: you add prompts, Visby runs them, and you see mentions, position, sentiment and which competitors showed up. Then it tells you what to do.
The 'tells you what to do next' angle is what separates it from a plain tracker. Several G2 reviewers say that turned AI visibility from a vague idea into a concrete content plan. One small operator called it a leveller against larger platforms.
What you actually get
From the official pricing page and the reviews, the feature set looks like this:
- Prompt tracking with brand mentions, citations, competing brands and visibility per prompt
- Competitor comparison and share of voice, including direct and indirect competitors
- Funnel-based tracking: awareness, consideration and decision stages
- AI referral traffic through Google Analytics, showing visits from ChatGPT, Gemini, Claude, Perplexity and Copilot, plus the landing pages
- Search Console and Bing connections
- Automated GEO task generation, content gap detection and content briefs
- Social proof tracking across Reddit, G2 and Trustpilot
- A report builder with PDF export, which agencies and consultants will like
A YouTube walkthrough by Alston Antony covers most of this live, including the report builder and the task manager. He is upfront that he uses an affiliate link and that parts of the product still need work.
What reviewers like
The praise is consistent. Onboarding is quick and clear. The dashboards are readable. One G2 reviewer saw where their brand sat against competitors within minutes. Another liked that the tool explains the data instead of just showing a score. Trustpilot reviewers say much the same: simple interface, actionable tasks, less guesswork.
A word of caution on the numbers. 4.8 stars across 109 G2 reviews and 68 Trustpilot reviews is a very uniform result, and 91% of G2 reviews are five stars. Vendor-invited reviews skew positive. Read the three-star-style complaints, since those tell you more.
What reviewers complain about
- Competitor detection needed manual cleanup. One reviewer said Visby miscategorized the brand and listed unrelated high-profile sites as competitors.
- The Reviews area showed zero results for one team, with no obvious way to point it at review pages.
- Traditional Chinese isn't supported.
None of these are dealbreakers. The first one matters most, though, because share of voice is only as good as the competitor set behind it.
Pricing and limits
All figures are from visby.ai/pricing. Check it before buying, since plans change.
| Plan | Price | Engines | Domains | Tracked prompts | Answers analyzed/month | Articles/month | Seats |
|---|---|---|---|---|---|---|---|
| Starter | $79/mo | Any 3 of 4 | 1 | 15 | 45 | 5 | 1 |
| Growth | $199/mo | Any 3 of 4 | 3 | 90 | 270 | 15 | 3 |
| Enterprise | Custom | Any 3 of 4 | 5+ | 250+ | 750+ | 25 | 10+ |
Every plan has a free trial of Starter, funnel tracking, automated task generation and social proof tracking. Enterprise adds a GEO deep-dive audit and a consultation call.
The quota math is the catch
One analyzed answer equals one tracked prompt run on one engine. Divide the quotas and every tier lands at 3 answers per prompt per month: 45 over 15 prompts, 270 over 90, 750 over 250. With 3 engines selected, that is about one run per prompt per engine per month. That's my arithmetic from the pricing page, not a number Visby states.
Why it matters: AI answers change from run to run. A single monthly sample per engine tells you whether you showed up once, not how often. A brand that appears in 40% of runs and one that appears in 5% can both look 'mentioned' or 'not mentioned' on any given month. You'd get a snapshot, not a rate.
The engine cap is the other limit. You choose 3 of 4 engines on every plan, even Enterprise, and Perplexity isn't offered. If your buyers use Perplexity or Copilot, or you want to compare ChatGPT and AI Overviews side by side, you'll need to drop one.
Why mention-only tracking is thin
Promptwatch Data's average sources per response report puts ChatGPT at about 5 cited sources per web-search answer, while Google AI Overviews and Perplexity cite about 10. It also says to measure AI Overviews and ChatGPT separately, because one doesn't predict the other. Copilot's average swung from under 2 to nearly 17 sources in a few weeks, which shows how unstable these numbers get.
Two things follow from that. With about 5 slots in a ChatGPT answer, whether you were cited at all is a blunt measure. You want to know which page was cited, which claim was attached to your brand, and where you landed in the source list. And since engines behave differently, a tool that limits engines and runs each prompt once a month leaves you guessing.
Other Promptwatch Data reports show how quickly the ground moves. Reddit's share of ChatGPT citations fell from about 4% to 0.5% on August 14, 2026, and ChatGPT started using the site: operator at scale on August 8, 2026. Monthly sampling won't catch changes like these when they happen.
What Visby doesn't cover
Based on the research I have, there's no mention of these in Visby's feature set:
- IP-verified AI crawler logs, to see which AI bots read your pages and hit errors
- Query fan-outs, showing the web searches AI runs behind each answer
- Prompt volumes and difficulty scores to pick which prompts matter
- Automated publishing to a CMS
I can't prove absence from a pricing page, so check with Visby on any of these that matter to you.
Who Visby suits
Good fit:
- A startup or solo founder who wants a first read on AI visibility for under $100 a month
- A small marketing team that values a ready-made task list over raw data
- A consultant who wants quick PDF reports for clients
Poor fit:
- Teams that need statistically meaningful tracking of many prompts across all major engines
- Anyone who needs Perplexity coverage alongside three other engines
- Enterprises with security review needs, or multi-language reporting in Traditional Chinese
- Teams who want technical diagnostics such as crawler access, not just visibility scores
Alternatives worth comparing
The category is crowded. A few that come up most often against basic trackers:
Promptwatch goes past monitoring. It tracks the real interfaces of ChatGPT, AI Overviews, AI Mode, Gemini, Perplexity, Copilot and more, adds IP-verified crawler logs and visitor analytics, and has a Content Agent that writes and publishes to Webflow, Framer or WordPress. Its Essential plan is $95/month for 50 prompts and 6,000 responses, versus Visby's $79 for 15 prompts and 45 answers. That is a far larger sampling budget. It is also rated 4.7/5 on G2 and is ISO/IEC 27001 certified.
Otterly.AI is an accessible entry-level tracker with mention tracking and basic sentiment. It's similar in spirit to Visby on the monitoring side.
Peec AI is monitoring-focused and doesn't try to optimize for you.
Profound is an enterprise option with a heavier price tag.
Radarkit has its own Visby review and covers a similar tracking and content-optimization space.
| Tool | Strength | Watch out for | Best for |
|---|---|---|---|
| Visby AI | Easy setup, task list, funnel view | 3-of-4 engine cap, thin sampling | Startups testing the waters |
| Promptwatch | Tracking plus crawler logs, traffic attribution and content publishing | More to learn than a basic tracker | Teams that want to fix visibility, not just watch it |
| Otterly.AI | Low-cost monitoring | Limited citation detail | Light monitoring |
| Peec AI | Clean monitoring | No optimization | Reporting-only needs |
| Profound | Enterprise depth | Price | Large brands |
For a wider shortlist, the GEO software directory at bestgeosoftware.com lists dozens of options by category.
How to test Visby without wasting a month
If you try the free trial, use it like a skeptic:
- Fix the competitor list on day one. Delete anything that isn't a real rival before reading share of voice.
- Choose engines on purpose. Pick the three your buyers use, not the default.
- Spend your 15 prompts on money prompts: comparisons and recommendations that map to the decision stage, not vanity queries.
- Connect Google Analytics and Search Console right away. AI referral traffic is the number that tells you whether visibility pays off.
- Run a manual check on a handful of prompts directly in ChatGPT and Gemini. If Visby's results disagree with what you see a few times in a row, take the sampling limits seriously.
- Judge the tasks by whether you'd do them anyway. A generic 'add schema' task is cheap. A gap tied to a specific prompt and a competitor is worth more.
Verdict
Visby does what it says: it gets a small team from zero to a working AI visibility dashboard quickly, and the task generation makes it feel less like a score and more like a plan. For $79 a month that's reasonable.
But it is basic mention tracking. About one run per prompt per engine per month, three engines out of four, and no sign of crawler or fan-out data means you see whether you appeared, not how reliably or why. If AI search is a side experiment, start there. If it's becoming a revenue channel, you'll outgrow it, and it's worth pricing out a platform with deeper sampling and execution tools before you commit your prompts and history to a tool that caps both.
For more tracking options, the AI rank tracking directory at ai-rank-tools.com is a good place to compare pricing side by side.