AI visibility now records the web searches Claude and ChatGPT run while answering your tracked prompts, and the pages each search returned.
GET /v1/visibility/queriesgroups those searches by phrase and by type, says how many runs searched each phrase, and lists the gaps: searches where a tracked competitor’s page came back in at least 2 runs and yours came back in none.- Each run from
GET /v1/visibility/runsnow carriesqueries, the searches behind that one answer in the order the engine ran them.queries: nullmeans the engine does not report its searches, which is always the case for Perplexity and Google AI Overviews. - The same data is available as the
traceten_visibility_queriesMCP tool and thetraceten visibility queriesCLI command. - A search phrase is generated by the model from your tracked prompt, sometimes reusing its wording. It is not a search any visitor or buyer ran.
The AI Crawlers page now tells you which pages AI crawlers fail to read, which pages they read to answer people, and whether that reading brings visitors.
- Needs attention lists pages where a verified AI crawler got a redirect or an error, with the company, how often it happened, when it was last seen and one line on what to do. A 401 or 403 is labelled Blocked, because a real crawler being refused usually means a firewall, CDN or WAF rule.
- An empty Needs attention list says why it is empty. If your install reports no response status, the panel says so instead of showing a clean result.
- Check pages requests your most-crawled pages and reports each one’s status, redirects,
noindex, and which AI crawlersrobots.txtblocks from it. On a Next.js middleware install, which cannot report a response status, this is how you check your pages. - What AI uses ranks the pages AI companies fetch to answer a live question, beside the AI visits and revenue each page brought. A page that is read often but never visited gets a Read, no visits tag. That is normal, because an assistant can answer from your page without sending anyone to it.
- The same data is available from three new endpoints,
GET /v1/ai-crawls/fix-list,/answer-mapand/reads-vs-visits, and as MCP tools and CLI commands. - The install guide now has a table showing which AI crawler adapters can report a response status.
Sources, referrers, AI crawlers, AI visibility and your own sites now show each website’s real icon.
- A website without a built-in logo shows its own favicon instead of a colored letter. Well-known brands keep their sharp built-in logo.
- Icons load from Traceten itself. Our server asks DuckDuckGo’s icon service for each website’s icon by its main domain (ycombinator.com for news.ycombinator.com), so loading an icon never sends your IP address to DuckDuckGo, and we do not store the icons.
- Sources with no logo of their own show an icon for what they are: Direct, Search, Social, Email, Paid, a website, an AI assistant, a crawler or a company.
- Logos sit directly in the row, without the box around them.
Traceten subscriptions are now sold through Dodo Payments.
- Dodo Payments is the seller named on your invoices and adds any sales tax, VAT or GST that applies at checkout. Prices are in US dollars. The charge on your card statement starts with
DODOPAY_. - An upgrade applies as soon as the payment is confirmed, usually within a minute, and it now starts a new billing period that day. You are charged the new plan’s price, minus credit for unused time on your old plan, plus any pay-as-you-go usage so far. Your monthly allowances restart with the new period.
- Downgrades and cancellations still take effect at the end of the period you have paid for, and you can undo either before then.
- If a renewal payment fails, you keep your plan while it is retried. If it stays unpaid, the account drops to Starter limits, with extra sites paused rather than deleted, until you update your card.
- The Payment section shows your card’s network, last four digits and expiry, and Update opens the Dodo Payments portal for your payment method and invoices.
Run
traceten ui to browse your numbers in the terminal, without typing a flag.- Pick a site from a list. The CLI remembers it and opens it straight away next time.
- The overview shows visitors, sessions, pageviews, AI visitors and AI sessions with their share, conversions, net revenue, bounce rate and active time, each compared with the previous period, beside a chart of sessions and AI sessions.
- Compact panels show your top AI assistants, how your traffic splits between AI and other sources, your top pages from AI traffic and your top sources. In a wide terminal they sit side by side; in a narrow one they stack.
- Press
[and]to move between the last 7, 14, 30 and 90 days,rto reload,sto switch site and?to see every key. - It uses the same credential as every other command, so run
traceten loginfirst. In a script, keep usingtraceten stats overview --json. - Every other command now prints tables at a terminal. Piped or redirected, output is the same JSON as before, and
--jsonstill prints compact JSON anywhere. - Higher API and MCP rate limits on every plan, so AI assistants and scripts that make many reads in a row are far less likely to see a
429. See Rate limits.
Connect Shopify with an app you create on your own store, so revenue attribution works the day you set it up, without waiting on an app review.
- Create a custom app in the Shopify Dev Dashboard with three read-only scopes, then paste your store domain, Client ID and Client Secret into Settings → Integrations.
- Traceten checks the credentials against your live store, registers the webhooks it needs, and imports your last 60 days of paid orders.
- Nothing is installed on your store. Traceten creates no script tag, theme file or app block, and asks for no write access.
- Your Client Secret is encrypted before it is stored, and only its last four characters can be read back.
The dashboard has a new, calmer layout that stays consistent from page to page, on a laptop and on a phone.
- Press S to switch sites and D to open the date picker. Switching sites keeps you on the page you were on.
- Goals and Funnels now have their own pages.
- Set up with AI gives you a ready-made prompt to paste into your coding assistant, for installing the snippet or connecting a revenue source.
- Sources, Revenue, Events, Conversions, Geography, AI crawlers, AI visibility and Settings were all rebuilt on the new layout. Sources gains All, AI, Normal and Unknown tabs, and Revenue gains a sales-by-channel chart.
Attributed revenue is now net of refunds, and Stripe Payment Links attribute without any code.
- A refund is subtracted from the source the original payment was credited to, on the day the refund was issued. A past day’s figure never changes after the fact.
- Revenue shows Sales (before refunds), Refunds and net Revenue side by side. A refunded order still counts as an order.
- Refunds arrive automatically from Stripe, Shopify, Lemon Squeezy, Polar and Paddle. Through the Payment API, send
refunded_amountas the running total refunded so far. - For a Stripe Payment Link, set the confirmation page to a page on your site with
session_id={CHECKOUT_SESSION_ID}appended. The snippet reads it and matches the payment to the visit. - A new Dodo Payments guide shows how to send payments to Traceten from Dodo’s own dashboard, with no server of yours involved.
The dashboard now counts people rather than visits, and your charts can carry notes that explain a spike.
- The new Visitors page has one row per person, with the source that brought them, what they spent, their time on site and when they were last seen. A customer who came back nine times is one row.
- Open a visitor to see every visit they made, a six-month activity calendar, and a day-by-day feed of the pages they viewed, what they clicked, the events they triggered and where they left.
- Filter visitors by country, city, device, OS, browser, channel, source, campaign, ref, page and AI assistant. The same data is available from the Visitors API.
- Pin a note to a day on the Traffic and revenue chart, such as “Launched pricing page”. Notes are shared with your team and can also be read and written over the API, the CLI and the MCP server.
- The Pages card gains an Exit link tab that ranks the domains visitors left for by clicking a link.
See where each visit ended, and trust that
track() calls wait for your consent banner.- Every visit now records the last page it was on and, when the visitor followed a link off your site, the domain they went to. We record the domain only, never the path or query string of the destination.
- Both appear in a visitor’s activity feed and on sessions returned by the API.
- When a consent banner (IAB TCF, OneTrust or Cookiebot) is still waiting for an answer,
traceten.track()calls are held in the browser. Nothing is sent and no cookie is written until the visitor accepts. If they decline, the held calls are discarded. - On the free Starter plan, AI crawl recording pauses when you reach your plan’s crawl allowance rather than running on, and resumes at your next billing period or as soon as you upgrade. Settings → Usage shows which state you are in.
Watch who is on your site right now on a live 3D globe.
- One marker is one visitor, over a rolling window of the last 12 hours. Visitors active in the last 5 minutes are highlighted, and the count appears as Online now on the Overview.
- A scrubber replays the last 12 hours, and a running activity feed shows pages viewed, goals completed, payments and exits as they happen.
- Select a visitor to filter the feed to that person. Names and avatars are generated and randomly assigned, not collected.
- Open the globe from Online now on the Overview or from Launch 3D globe on the Geography page.
- Traceten now suggests a cookie domain derived from your site’s address. Save it, and every snippet the install page gives you carries it, so a visitor who moves between your subdomains stays one visitor.
AI visibility, now in Beta, shows how often AI answer engines name you and cite your pages when asked the questions your buyers ask.See AI visibility, AI answer engines and Visibility API.
- Track your own prompts across Perplexity and Google AI Overviews, on every plan. Traceten asks them on a schedule and stores every answer so you can read it verbatim.
- Add Claude and ChatGPT with the engine coverage add-on. Both are measured through each vendor’s API with web search turned on, and the docs explain how that differs from the consumer apps.
- Three separate numbers, never blended into one score: presence rate, citation rate and share of voice. Every figure is shown as a range.
- The Citations tab shows which pages the engines linked to, yours and everyone else’s, joined to the AI crawler fetches and the AI-referred revenue on your own cited pages.
- Confirm your brand, add competitors, and let Suggest propose prompts from your own pages. Everything is also available over the API and the MCP server.
AI visibility is in Beta. Read How AI visibility is measured before you act on a number.
Run Traceten from your terminal, and connect AI apps by signing in instead of pasting an API key.
- The Traceten CLI runs any API endpoint from a shell and returns JSON you can pipe into
jq. Its commands are generated from the same spec as the API reference, so the two always match. traceten loginopens your browser. Approve the CLI, pick your organization, and you are signed in, with no key to create.- In CI, pipe an API key into
traceten login. It never opens a browser. - Claude, Claude Code, Codex, ChatGPT and VS Code can now connect to the MCP server by signing in to your Traceten account. The connection sees only what you can see.
- A signed-in session can read your numbers and edit goals, funnels and sites, but it can never mint an API key or a secret. Removing someone from your organization cuts off their connections within about 30 seconds.
Attribute revenue from Lemon Squeezy, Polar and Paddle to the AI source that sent the buyer.
- Connect each one by pasting an API key (an organization access token for Polar) and your store, organization or seller ID. There is no OAuth redirect.
- Traceten creates the webhook on your account for you, so there is nothing to configure on the provider’s side.
- Your last 90 days of paid orders are imported when you connect.
- Refunds from all three are subtracted from the source that earned the original payment.
Attribute revenue from any payment processor, including ones Traceten has no built-in integration for.
- Send each payment to
POST /v1/server/paymentswith an API key that has theingest:writepermission. - Identify the customer by their Traceten visitor ID or the email address they paid with, and the payment is matched back to the visit that produced it.
- Payments without either are still recorded, as unattributed revenue.
- Use it for Podia, Kajabi, Ghost, your own checkout, or anything else that can make an HTTP request.
Ask an AI assistant about your AI traffic in plain English, in the conversation you are already having.
- Traceten runs a remote MCP server at
https://api.traceten.com/mcp. There is nothing to install or run locally. - It launched with 59 tools that read your numbers and manage your goals, funnels, sites and webhooks.
- Every tool call goes through the same API as a direct request, so the key’s permissions and site scope apply to everything your assistant does.
- Tools that delete something, issue a credential or change how measurement works show a preview first, and act only after you confirm.
Read your analytics and manage your account programmatically, with API keys you scope to exactly what they need.
- Give each key its own permissions, such as reading analytics, editing configuration, or issuing credentials, and limit it to one site or all of them. New keys can only read until you grant more.
- Read stats, breakdowns, sessions, revenue, AI detection results and AI crawler data over HTTPS.
- Create sites, mint API keys and register webhooks through the management API.
- Rate-limited endpoints on
api.traceten.comreport your remaining headroom inx-ratelimit-*headers on every response, and a429tells you how long to wait.
Take payments in other currencies and connect your payment provider in one paste.
- Payments are stored in the currency you were paid in, and converted to USD for reporting at the European Central Bank reference rate for the day of the payment.
- The exchange rate used is saved with every payment, so any total can be traced back to the numbers behind it.
- A payment in a currency Traceten cannot convert is rejected with an error naming the currency. It is never recorded as zero revenue.
- Each site has a display currency for its totals, which defaults to your payment provider’s account currency when you connect one.
- Connect Stripe by pasting a restricted API key. Traceten creates the webhook endpoint on your Stripe account for you, with no OAuth redirect and nothing to configure.
Invite your team and control who can do what.
- Invite people by email from Settings → Team. Each is an admin or a member.
- Admins manage billing, sites, API keys, crawl tokens and integrations. Members see all analytics and can change site settings and webhooks.
- Limit a person to specific sites from the Site access tab. Everything they reach, including any AI app they connect, follows that limit.
- The number of seats depends on your plan.
Count the actions that matter and see where visitors from each AI assistant drop off.
- Fire a goal with an HTML attribute and no JavaScript:
data-traceten-goalon a click, ordata-traceten-scrollwhen a section comes into view.traceten.track()and the server SDKgoal()work too. - Any new event name becomes a goal automatically. There is nothing to create in advance.
- Build a funnel of 2 to 8 steps from page visits and goals, with a conversion window from 1 hour to 90 days. A step can be negated, for visitors who did not do something.
- Each funnel step shows visitors, drop-off, revenue from the visitors who got that far, and the sources and countries they came from.
Events and AI crawler hits are now metered separately, each with its own allowance.
- AI crawler hits never count against your event allowance.
- Settings → Usage shows a card for each meter, counted over your billing period.
- Pay-as-you-go overage has a separate switch for each meter, so turning one on never turns on the other.
- On a paid plan, going past your crawl allowance never blocks a crawl or loses a count.
See which campaign drove traffic and revenue, not only which channel.
- Traceten now captures the
utm_campaignandrefparameters from the landing URL. - Break traffic down by campaign on the Overview’s Traffic sources card, and filter visitors by campaign or ref.
- Both are for reporting only. They never change how a visit’s source is detected, so nobody can move your AI numbers by writing a link.
Change plans from inside the dashboard, and undo a deletion for 30 days.
- Compare every plan side by side and switch from Settings → Billing. Upgrades apply immediately, and downgrades apply at the end of your billing period.
- Before you confirm, the dialog tells you what changes, what it costs and when it takes effect.
- Deleting a site or your account now stops tracking straight away but erases nothing for 30 days, so you can cancel. After that, erasure is permanent.
See where in the world your visitors come from.
- The Geography page shades a world map by visitors per country, with the same countries listed beside it, busiest first.
- Select a country for its visitors, pageviews and revenue, the sources that sent them, the devices they used and the pages they read.
- Small countries such as Singapore and Malta are drawn as dots, so they never disappear from the map.
- The same map appears on the Overview’s Places card.
See which AI systems read your pages, why they read them, and whether they are who they claim to be.
- AI crawlers do not run JavaScript, so the snippet cannot see them. Add the
@traceten/ai-crawlmiddleware to your server (Next.js, Express, Hono, Cloudflare Workers or Cloudflare Pages) with a per-site crawl token. - Each crawl is sorted by purpose: answering a live question, building a search index, or collecting training data.
- Traceten checks each crawl against the IP ranges and reverse DNS the vendor publishes, and marks it verified, unverified or spoofed.
- The new AI Crawlers page and an Overview card show who crawled, which pages, and when. The same data is available from the API.
- Browse every tracked AI crawler in the public directory at traceten.com/crawlers.