toseo.todayHome

Last updated October 28, 2026

Privacy policy

toseo.today turns a website's SEO data into daily to-do lists. This page explains what data the app uses, where it is kept, and how to remove it. In short: we only read your data to build your task lists and Insights, we don't sell it or use it for advertising, and you can disconnect at any time.

Google data (Search Console and Analytics)

When you choose "Sign in with Google", you allow toseo.today to:

  • See your Search Console data (webmasters.readonly): the properties you can access, and the searches, clicks, impressions and positions for the property you add as a project.
  • See your Google Analytics data (analytics.readonly): the GA4 properties you can access, and organic landing-page engagement, key events and revenue for the property you add, plus visits from AI assistants (sessions and key events by AI source and landing page).
  • See your email address, to show which Google account is connected.

Access is read-only: toseo.today can't change anything in your Google accounts.

How we use it:

  • The data is used only to show you Insights (pages losing traffic, searches you could win), to rank your SEO tasks by the traffic they could bring, to list searches that are new for your site this week in Keyword Research, to show where your site ranks for each search (and, for the searches your team tracks, day by day) in Keyword Research's Rankings tab, to show which AI assistants send you visitors on the AI visibility page, to build the Topical map (each page's clicks, impressions and position over three months, for all countries and the US), and to write content briefs (the searches a page shows for, with their clicks, impressions and positions, for all countries and the US, or which of the site's pages show for a search).
  • It is fetched when you open the app, processed on our server, and kept in the server's memory for up to one hour so pages load quickly. The data itself isn't stored in a database, with two exceptions: the Topical map keeps each of the site's pages with its clicks, impressions, average position and top search for the three months it was built from, as described under "Topical map" below; and a content brief keeps the searches of the page it plans (up to 100, with their clicks, impressions and positions), as described under "Content briefs" below. If you add an insight to your plan, the task it creates (which can mention a search and its clicks) is saved with your team's progress, as described below.
  • It isn't sold, used for advertising, or used to train AI models. The only company it goes to is Anthropic: the search terms and page addresses described under "Link-building ideas" below, and for a content brief, the searches of the page it plans with their impressions and positions (for a new page, which of the site's pages already show for the search, with theirs), as described under "Content briefs". Nobody at toseo.today reads it unless you ask us to help with a problem.
  • To keep you signed in, Google gives us a refresh token. It is encrypted and kept in a cookie on your device (toseo_vault), not on our servers, for up to 400 days unless you disconnect.

toseo.today's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.

Ahrefs data and API keys

If you connect Ahrefs, your API key is encrypted and kept in the same cookie on your device. It is used only on our server to read your Site Audit projects, issues and affected pages. Audit results are cached until a newer crawl is available, so we don't use more of your Ahrefs API units than needed. When someone on your team asks for it on the Connect page, the key also reads the Site Audit's list of the site's indexable pages (address, title, main heading, word count and how many of the site's pages link to each), saved as the project's page list for the Topical map, as described under site crawls below. When someone on your team runs Keyword Research, the key is also used to read public Ahrefs data about the project's site and its competitors (rankings, keyword metrics, and who ranks in Google for a keyword). The resulting report is saved in team storage until the next run or until the project is removed. When the Topical map is updated, the key also reads the US search volume of the map's gap searches (searches made from the site's topics, such as "is coinbase safe"), of the focus keywords your team lists and of the searches competitors' pages rank for that no page of the site covers, the site's organic competitors, and up to 200 top pages of up to four competitors (each page's address, the search it gets the most traffic from, and how many searches it ranks for), which are compared with the site's own pages to show what competitors have that the site doesn't; these are kept with the map for 30 days. When someone on your team creates a content brief, the key also reads, for the brief's search: who ranks in Google's top 10 in the US (with its People Also Ask questions), and in one other country when the US isn't the whole story; the search's difficulty and monthly searches, and by country when needed; and the questions and related searches that contain it. A brief uses at most 2,500 units. To keep each feature within the team's monthly limits, the app also counts the Ahrefs API units each feature uses, per month, in team storage. The count is stored with a one-way fingerprint of the key, never the key itself, and kept for about a year.

Keyword Research's trending tab sends the project's topic words (for example "bitcoin", never your Search Console figures) to Google Trends and Google News to find rising searches and recent headlines, and reads CoinGecko's public list of trending coins for crypto sites. Results are kept in the server's memory for up to one hour.

AI visibility

When someone on your team checks AI visibility, toseo.today uses the project's Ahrefs key to read public Ahrefs data: how often AI answers (Google's AI Overviews and AI Mode, ChatGPT, Perplexity, Gemini and Copilot) cite the project's site and the competitors you list, which Google searches show an AI Overview that cites the site or doesn't, and which pages those AI Overviews link to. To check what AI crawlers can read, our server reads the site's robots.txt, a few of its pages, and up to three pages each checked AI Overview links to, the way a search engine crawler would. It says it is toseo.today; for your own site only, a page that turns it away is tried once more as a browser. Only the findings are kept (counts, searches, page addresses and the result of each check), in team storage, until the next check or until the project is removed. No page content, and nothing from Search Console or Analytics, is sent anywhere else.

If you import an AI report exported from Search Console or Bing Webmaster Tools, the file is read in your browser, and only its page addresses, queries and counts (up to 300 of each) are saved in team storage, until you remove them or the project.

When someone on your team asks the AI assistants, each question they typed (up to 10) is sent, on its own, to the assistants set up on this deployment with toseo.today's API keys: OpenAI (ChatGPT, with web search and with storage turned off), Google (Gemini, with Google Search), Perplexity (Sonar) and Anthropic (Claude, with web search). Nothing else goes with it: not the site, the competitors, or any Search Console or Analytics data. Each company handles the question under its API terms; OpenAI, Perplexity and Anthropic don't train on API data by default, while Google may use questions sent with a free-tier Gemini key to improve its products. The answers (up to 6,000 characters each) and the pages they cite are saved in team storage (or the server's memory without it) for the latest run, with the counts of earlier runs, until the project is removed.

Measure

When someone marks a task done (or logs a change by hand), toseo.today saves a short record of it: the day, the task's type and title, the page address, the search it's about, and the old and suggested value shown in the task (for example the old title). To measure what the change did, our server reads that page (the way a search engine crawler would, as for AI visibility) to check the change is live, asks Search Console's URL Inspection when Google last visited up to 20 changed pages, and reads daily Search Console numbers for the changed pages, their searches and the whole site. The records, with the result of each page check, Google's last visit and each final result (click, impression and position totals), are kept in team storage (or in your browser without it) until the task is marked as not done, the logged change is removed, or the project is removed. The daily numbers themselves are only held in the server's memory for up to an hour.

Topical map

The Topical map is built from the project's page list (from a site crawl or its Ahrefs Site Audit), its Search Console numbers per page (three months, for all countries and the US) and the Ahrefs data described above. When a project has no page list yet, opening the Map tab crawls the site from your browser tab the way a crawl from the Connect page does (described under site crawls below), but keeps only the page list. Until your team saves its own strategy, our server works out the site's profile (its central entity, sections, the attributes of each entity and the rules that place pages) from the page list's addresses, titles and main headings, again for each newer page list. The map, with each page's figures, the gaps, what competitors have and the recommendations, is saved in team storage (or the server's memory without it) until the next update or until the project is removed. So is the profile and the strategy: the one worked out from the pages, with how it was worked out, or the one your team saves on the map's Strategy panel (the site's central entity, sections, attributes, topic rules, focus clusters and keywords, competitors and update settings). When someone approves the map as the site's silo, a copy is saved with the name they type and the date, so later updates can show what changed since.

Claude can propose the site's central entity, source context and search intent. That request carries the same details as the link-building ideas below: the project's domain and brand names and, when Search Console is connected, its 15 top non-branded searches and 10 most visited page addresses, with no figures; and what the site's pages show (its sections, up to 12 of the names its pages cover, the attributes and the main list page). When the deployment has an Anthropic API key, it is sent automatically each time the profile is worked out from a new page list, until your team saves its own strategy, and the answer goes into the profile, which your team can change. It is also sent when someone on your team asks for a proposal on the Strategy panel, through the API or by copy and paste into their own Claude account; then the answer only fills in the form until they save it.

Pages the map can't place are sent to Claude, Anthropic's AI model, which proposes a topic for each: only their addresses, titles and main headings go with the request, together with the site's profile (its main subject, sections, entities and topics), and no Search Console or Analytics figures. Like the link-building ideas below, it goes to Anthropic's API automatically when the deployment has an Anthropic API key, or otherwise by copy and paste into a team member's own Claude account. The proposals are saved with the map until your team approves or turns them down; approved ones become part of the project's strategy.

Content briefs

A content brief is made only when someone on your team creates one, from a Keyword Research idea, a gap on the Topical map or a typed search. It uses the project's Topical map and page list (where the page fits and which pages to link), Search Console (the page's searches with their clicks, impressions and positions, for all countries and the US; for a new page, which of the site's pages already show for the search) and the Ahrefs data described above.

toseo.today then opens the pages that rank for the search (at most five editorial pages and three of the brand's own pages, plus your page when the brief updates it) and keeps only their outlines: their headings, how many tables, lists and images they have, and their word counts, never their text. A site that turns automated visits away isn't read.

Claude, Anthropic's AI model, writes the brief's draft, the same two ways as the link-building ideas below: automatically through Anthropic's API when the deployment has an Anthropic API key, or by copy and paste into a team member's own Claude account. The request carries the brief's search, the site's address and the code you gave for it (if any), where the page fits and the pages to link, the searches with their Search Console impressions and positions (for a new page, which of the site's pages already show for the search, with theirs; clicks are left out), the questions people ask, and the compared pages' outlines. Through the API, Claude also searches the web for the official page that states each value the brief needs. Every source page Claude proposes is then opened to check that it mentions that value; its text is read for that and not kept.

The brief, with what its steps collected (including the searches of the page it plans, up to 100 with their figures) and Claude's draft, is saved in team storage (or the server's memory without it) until someone deletes it on its page or the project is removed, together with the dates it was approved and marked as sent to your writer. Approving a brief adds its task to your team's plan.

A brief has a part for the writer and a part for review. Only the part for the writer can be copied for Google Docs or downloaded as a Markdown file: it holds the outline and instructions, the values to fill in with their sources, and the code you gave; the search data and the reasons behind the brief stay in the part for review. The copy and the file are made in your browser from the brief on the page; toseo.today doesn't send the brief to Google Docs, your writer or anyone else. It leaves the app only when someone on your team pastes or sends it.

Link-building ideas (Claude by Anthropic)

The off-page Tactics and Digital PR tabs are written by Claude, Anthropic's AI model, with web search. To tailor them, the request includes the project's domain and brand names and, when Search Console is connected, its 15 top non-branded searches and 10 most visited page addresses (no click counts or other figures), plus the titles of the tactics it already has. This happens when a project first opens these tabs and then once a week. It reaches Anthropic in one of two ways:

  • Automatically, when the deployment has an Anthropic API key: toseo.today sends the request to Anthropic's API. Under Anthropic's commercial terms, API data isn't used to train its models.
  • By copy and paste otherwise: the app shows the request as a prompt, and a member of your team pastes it into their own Claude account (claude.ai) and pastes Claude's answer back. That chat is covered by the terms and privacy settings of their Claude account; whether it may be used to improve Claude depends on the "Help improve Claude" setting there.

The resulting tactics and ideas are saved in team storage so your team sees the same list.

Site crawls and crawl files you import

When someone on your team crawls a project's site from the Connect page, our server fetches the site's public pages a few at a time, the way a search engine crawler would: it reads robots.txt and the sitemaps first, follows the rules robots.txt sets for crawlers, and says it is toseo.today (ToseoBot). Only if the site's firewall turns it away is it tried as a browser. To rank the fixes it also reads the project's Search Console clicks per page, with the search that brings each page the most clicks, for the last 28 days. Your browser tab runs the crawl and receives each page's findings (status code, title, meta description, headings, word count, canonical and robots tags, and the links between the site's pages); page text isn't kept. After the crawl, up to 50 of the site's page addresses are sent to Search Console's URL Inspection (with the project's Google connection) to see whether Google has indexed them, and up to 6 to Google's PageSpeed Insights, which loads them and returns Core Web Vitals from real Chrome visitors and a lab test. When the crawl ends, the audit is kept, the same way as for a crawl file below, including each checked page's index status, Google's last crawl date and its speed figures. So is the site's page list for the Topical map: each page's address, title, main heading, word count and whether it can be indexed, which of the site's pages link to which, and how many pages link to each (news sections are left out and only counted). It is saved in team storage (or the server's memory without it) until the site is crawled again or the project is removed. When the Topical map finds no page list for a project, opening its Map tab runs the same crawl from your browser tab (up to 1,000 pages, news sections last) and keeps only the page list, not an audit.

Crawl exports (Screaming Frog, Sitebulb, Ahrefs or other CSV files, or a Screaming Frog issues report in Excel) are read in your browser. Only the results are kept: the issues found and, for each issue, up to 50 affected pages with their address, title, meta description, heading, status code, links and traffic figures from the file (or, for a crawl, from Search Console), plus a list of up to 1,000 affected addresses with their status code and traffic. From an issues report, each issue's description and fix, and your team's To-do and Note entries, are kept too. With team storage (below) they are saved there so your team sees the same tasks; otherwise they stay in your browser's local storage.

Your projects and progress

Projects (site name, Search Console property, GA4 property ID, Ahrefs project ID, brand terms) and progress (tasks done or moved, the workload, issues moved to Ignore, planned and dismissed insights, outreach statuses, the searches tracked in Rankings, and Measure's records of changes) are saved in toseo.today's team storage, a database hosted by Upstash (Redis), so everyone on your team works from the same lists. On a deployment without team storage they are saved in your browser instead (projects in the encrypted cookie, progress in local storage). Google sign-ins and Ahrefs keys are never put in team storage.

Cookies

The project you have selected is remembered in a cookie (toseo_project). During Google sign-in, a short-lived cookie (toseo_oauth_state, 10 minutes) protects the sign-in from forgery. We don't use advertising or tracking cookies.

Hosting

toseo.today runs on Vercel, which keeps standard request logs (such as IP address, page requested and time) to operate and secure the service. Requests to Google, Ahrefs, CoinGecko and the websites checked for AI visibility and Measure are made from Vercel's servers. Team storage is hosted by Upstash.

Removing your data

  • Disconnect Google or Remove key (Ahrefs) on the Connect page. Disconnecting Google also asks Google to revoke toseo.today's access. You can also revoke it yourself at myaccount.google.com/permissions.
  • Remove a project on the Connect page, which also deletes its progress, imported crawl, page list, Topical map and its strategy, approved silo, content briefs, link-building ideas and keyword research from team storage (or from your browser). A single brief can be deleted on its own page.
  • Clearing this site's cookies and site data in your browser removes everything toseo.today keeps on your device.
  • To have everything about your team deleted from team storage, email us (below).

Children

toseo.today is a tool for website owners and isn't meant for children under 16.

Changes and contact

If this policy changes, we'll update the date at the top of this page. Questions or requests: hello@toseo.today.

toseo.today

Tick it off. Watch it climb.

TourDemoPrivacyTermsContact
© 2026 toseo.today