The AI visibility tool boom: 400+ public traces in two years
The first version of this research was too strict. It only plotted products with a clean launch date. This version also uses first public traces: directory appearance, GitHub repository creation windows, Product Hunt records and company announcements. That gives us a much better view of how fast the category is spreading.
public traces
Broad research corpus
commercial tools
Temso and LLMrefs directories
verified launches
Day or month precision
tools
Elmo feature-comparison directory
There are two useful counts, not one
If the question is how many recognizable products exist, the safest answer is 200+. LLMrefs maintains a directory of more than 200 AEO, GEO, LLMO and AI-search tools. Temso independently maps more than 200 platforms. Elmo uses a tighter feature-comparison definition and currently lists 106.
If the question is how much tooling is being built around AI visibility, the number is already larger. Our wider research corpus passes 400 public traces once we include commercial products, open-source trackers, graders, WordPress plugins, browser tools, MCP utilities, APIs and serious public prototypes.
We do not call all 400 separate companies. Some public projects are experiments, some are product components, and some may overlap with a commercial tool. The broader count is useful for showing development activity. The 200+ commercial count is the better number for market sizing.
How we assign rough dates
An exact launch announcement is still the best evidence. We use that whenever it exists. For the long tail, we now accept the earliest public trace we can defend. A Product Hunt launch or official release gets day precision. A company launch month or first indexed appearance gets month precision. GitHub projects use their repository-creation window. When only a wider window is available, the chart treats it as an estimate rather than inventing an exact day.
That matters because founding date, incorporation date, beta date and public launch are often different things. The pale series below is therefore an estimate of visible ecosystem growth. The dark line is the smaller set where we checked the launch evidence one product at a time.
Broader first-public-trace timeline
From a handful to 400+ public traces
The pale area is the broader research corpus, using launch dates, directory appearance and GitHub creation windows. The dark line is the 46 launches dated individually.
Selected markers
AisoMar 2025How the category formed
2024: a few pure-play products
Profound is the first dated launch in our verified sample, on August 12, 2024. Otterly's company history places its initial product in October. SnowSEO appeared on Product Hunt later that month. Scrunch launched in November, according to TechCrunch's account of the company.
Spring 2025: the market becomes obvious
Peec launched in February. Semrush announced its AI Toolkit on March 4, and Ahrefs put Brand Radar into its March 5 product update. Aiso started that month. Superlines followed on March 19, then BrightEdge launched AI Catalyst in April.
Late 2025 and 2026: development activity explodes
Q2 2025 was the fastest quarter in our individually dated sample, with 10 launches. But the GitHub creation-window searches show a much bigger long tail from the second half of 2025 onward. By 2026, new trackers, auditors, APIs, local-search variants, plugins and open-source implementations are appearing constantly. Our verified sample contains26 launches from 2025, while the wider public-trace corpus grows several times faster.
The common product loop
Most tools still do the same four jobs
Choose prompts
Start with a prompt or topic set, typed by the customer, generated from keywords, inferred from competitors, or collected from real conversations.
Run AI answers
Send those prompts to ChatGPT, Gemini, Perplexity, Claude, Google AI surfaces, or a subset of them, then repeat the test on a schedule.
Create visibility metrics
Parse the answers for mentions, citations, position, sentiment, competitors, share of voice, and recurring source domains.
Suggest the next action
Add an audit, content brief, workflow, or distribution task. Newer products increasingly connect monitoring to execution.
Where the products actually differ
A feature checklist makes the market look more uniform than it is. These choices change what the number on the dashboard means:
- Where the prompt set comes from: synthetic prompts, search data, customer research, or real AI conversations.
- How the answer is collected: an API, a search endpoint, or the consumer interface a person actually uses.
- How much sampling is done across models, regions, languages, dates, and repeated runs.
- Whether the raw answer and source trail remain available for inspection.
- Whether the product stops at measurement or helps a team publish, distribute, and test changes.
The 46-product verified subset
The source register below remains the high-confidence layer. These are products or named modules where we found a repeatable AI-answer monitoring function and a sufficiently precise public date. Official release posts come first. Product Hunt and credible reporting fill gaps. Month-only dates are placed on the 15th and labelled as month precision.
Products can be real and still be absent from this smaller list. That no longer removes them from the main market story. It only means their timestamp belongs in the rough public-trace layer rather than the exact-launch layer.
All 46 individually dated launches
20244 launchesOpenClose
OtterlyAI
Company says the initial product launched in October 2024; plotted at mid-month.
Scrunch AI
Reporting places the product launch in November 2024; plotted at mid-month.
202526 launchesOpenClose
Peec AI
Company materials place the launch in February 2025; plotted at mid-month.
Aiso
Aiso started in March 2025. Its first Product Hunt launch followed on May 21, 2025.
202616 launchesOpenClose
The dashboard is now the easy part
Aiso combines controlled prompt tracking with real, consent-based AI conversations and the search fan-outs behind answers. That helps separate a neat score from the demand and evidence that produced it.