Privacy Policy

How Paradigma handles personal data and repository data. Effective October 13, 2026. Updated October 18, 2026.

“Paradigma”, “we” and “us” mean Alisson Ayres, an individual based in Brazil, who operates the Paradigma service at paradigma.click and is the controller of the account data described here. For repository, website and analytics data you connect, we process it on your behalf to provide the service. Contact: support@paradigma.click.

Account data

You sign in with GitHub, Google or an email code through Clerk, our authentication provider. We receive your name, avatar and email address, and keep a copy of your name and avatar to show in the dashboard. GitHub is optional; Paradigma reads repositories only through it: when you connect it, Clerk keeps the GitHub token and gives it to us when a check runs, to confirm your current access to installations. A session cookie keeps you signed in.

Repository data

The GitHub App asks for read access to metadata, deployments and organization members, read and write access to contents, pull requests and checks, and read access to email addresses. Code is analyzed in memory and never executed. We store findings, file paths, affected routes and the specific metadata values needed to explain a finding, not complete repository files. Installation tokens are created per job and not stored.

Website data

For sites you verify, our crawler reads public pages, robots.txt and same-site sitemaps. We keep page URLs, status codes, titles, descriptions, canonical and robots directives, counts and a content hash. Fetched HTML is discarded after analysis. See ParadigmaBot.

Connected services and API keys

If you connect Google Search Console, Google Analytics, Bing Webmaster Tools, Plausible, PostHog or Slack, we store the credentials encrypted and keep the metrics we read (such as clicks, impressions and AI referral counts). Google access is read-only: Search Console with the webmasters.readonly scope and Google Analytics with analytics.readonly, for the one property you choose. On Pro and Agency, AI visibility samples run on our own OpenRouter and DataForSEO accounts: the prompts, built from your Search Console queries, are sent to those providers to get answers. AI provider keys you add instead (OpenAI, Anthropic, Google Gemini, Perplexity, DataForSEO) are encrypted at rest and used only inside your sampling jobs. Either way, we store the prompt, a bounded answer, citations, model label and usage.

Paradigma agent

AI visibility sampling never sends repository source to a model. If your organization turns on the Paradigma agent, it writes fixes on our servers. For each fix it reads the files that fix needs, a bounded part of the repository and never secret files such as environment or key files, through the GitHub App, in the memory of one background job, and sends them to an AI model: the model provider never trains on them and deletes them after a limited retention period. It never runs your code. It commits the fix to a new paradigma/ branch and opens a pull request. We keep the pull request link, the changed file paths, line counts, the commit SHA, the model, token counts and cost, never file contents. Hosted fixes are opt-in per organization. The Paradigma agent can also run in your own GitHub Actions. By default it then uses the AI model Paradigma provides, under the same opt-in: each model request goes from your GitHub runner to our servers, which read it in memory and pass it to the same AI model, never storing or logging it; we keep the model, token counts and cost. If you turn on your own model key for a repository, the runner calls your AI provider directly under your provider account and its terms, and those model requests never reach our servers. In both cases Paradigma reads the files its checks need through the GitHub App and receives the files a fix changed, only to check them in memory, never storing them.

Billing data

Stripe collects and holds your payment details. We store your Stripe customer and subscription ids, plan and status, not card numbers.

Usage data

We record product events such as dashboard page views and check results to improve Paradigma. These contain ids and counts, never code, file paths or repository names. Our servers also keep standard request logs.

Visits to this website

When you visit paradigma.click, Vercel Web Analytics counts page views without cookies. The live site also uses Google Analytics 4 to measure visits: Google receives your IP address, browser and device details, the pages you view and the page that sent you, and Analytics sets cookies. Google signals and ad personalization are turned off. In the European Economic Area, the United Kingdom and Switzerland, analytics storage is off by default, so Analytics sets no cookies there. You can opt out with the Google Analytics opt-out browser add-on or by blocking cookies in your browser.

Paradigma'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. We use Google Search Console and Google Analytics data only to show you your own site's search and traffic metrics next to its findings. We do not sell it, use it for advertising, share it with third parties except our processors to run the service, or use it to train AI models, and people at Paradigma read it only with your permission, for security, or when the law requires. You can revoke access at any time in your Google Account permissions.

  • to run checks, crawls, AI visibility samples and fixes you ask for;
  • to show results in the dashboard, MCP and your alerts;
  • to bill you and manage your subscription;
  • to keep the service secure and fix problems;
  • to contact you about your account and important changes.

We do not sell personal data, show ads, or use your code or data to train AI models. Legal bases where applicable: performance of our contract with you, our legitimate interest in running and securing the service, and your consent where required.

We use these providers to run Paradigma:

  • Vercel: hosting, functions and background jobs
  • Supabase: Postgres database
  • Clerk: sign-in, sessions and the GitHub token for repository access
  • Stripe: payments, invoices and the billing portal
  • GitHub: repository access you grant
  • PostHog: product usage events (ids and counts, no code or repo names)
  • OpenRouter: AI answers for the AI visibility included in Pro and Agency (prompts only), and AI model access for the Paradigma agent under zero data retention, when we route fixes through it
  • Vercel AI Gateway: AI model access for the Paradigma agent under zero data retention, when we route fixes through it
  • Anthropic (Claude models), served by Anthropic or by Google Cloud Vertex AI or Amazon Bedrock: the AI model that writes Paradigma agent fixes, from the files a fix needs, through Anthropic's API, which never trains on them
  • DataForSEO: Google AI Overviews and AI Mode results for the AI visibility included in Pro and Agency (search queries only)
  • Google: Google Analytics for visits to paradigma.click

Services you connect yourself (Google, Slack, AI providers and others) receive requests made with your credentials under their own privacy terms. Data may be processed in the United States and other countries where these providers operate; where required, these transfers rely on the providers’ standard contractual clauses.

We keep account and project data while your account is active. Uninstalling the GitHub App stops our access to your repositories; stored findings and history remain until you ask us to delete them, which we do within 30 days. Removing an AI provider key or the Slack connection deletes that stored credential. Server logs are kept for up to 30 days.

Depending on where you live, you may have the right to access, correct, export or delete your personal data, and to object to or restrict its processing. Email support@paradigma.click and we will respond within 30 days. You may also complain to your local data protection authority.

Secrets and third-party credentials are encrypted at rest. Dashboard access follows your current GitHub installation access, checked on every request. The crawler only fetches HTTPS URLs on your verified domain that resolve to public addresses. No system is perfectly secure; if a breach affects your data, we will notify you as the law requires.

We will post updates here and, for material changes, notify you by email or in the dashboard. Questions or requests: support@paradigma.click. See also Terms of Service and Data handling.