Security and data handling
We only read your public pages.
RulerIQ reads the same public pages anyone can visit. We don’t sit in your CDN path, read your visitor logs or ask for write access to your CMS. Secrets you connect (a Linear key, or your own AI provider keys on own-key plans) are encrypted with AES-256-GCM using a key held outside the database, shown only as their last 4 characters, and used only for your checks (see the docs). We store scan results, the emails and domains you submit, IP addresses for rate limiting and, when enabled, analytics events (see Privacy). We hold no security certifications of our own.
Public pages only
We fetch a short list of public URLs on the domain you submit (the homepage, /pricing, /robots.txt, /llms.txt and sitemaps), honoring the site’s robots.txt. No visitor logs, cookies or internal documents.
Official provider APIs
We query OpenAI, Perplexity, Google Gemini and Anthropic through their commercial developer APIs, not scraped consumer apps. Prompts are built from public category and brand names.
No access to your site
No CDN proxy and no CMS write tokens. Fixes are inspectable FAQ HTML, JSON-LD and /llms.txt snippets you paste yourself. Search Console, if you connect it, is read-only.
SSRF-safe fetching
Every outbound fetch of a submitted URL pins the resolved IP and blocks private, loopback, link-local and CGNAT ranges, with redirects re-checked at each hop.
Subprocessors
The services that may handle data on our behalf. “When enabled” means the integration is built but not yet switched on for visitors. RulerIQ does not store card numbers, and the providers below hold their own certifications, not RulerIQ.
| Provider | Role | Data handled | Status |
|---|---|---|---|
| Fly.io (iad region) | Hosting for the web app and background workers | Public scan requests, application logs | Active |
| Fly Postgres | Database | Scan results, emails and domains you submit, IP addresses for rate limiting | Active |
| OpenAI, Anthropic, Google, Perplexity | LLM processors (commercial APIs) | Synthetic buyer prompts built from public category and brand names | Active |
| DeepSeek | LLM processor (commercial API) | Text from the scanned site's homepage, plus engine answers, cited URLs and site facts when tracing a wrong price | Not used at launch |
| Cloudflare Turnstile | Bot protection on the grader | Challenge token and request metadata | Active |
| Clerk | Accounts and sign-in | Account email and session | When enabled |
| Resend | Transactional email | Email address, report content | When enabled |
| PostHog | Product analytics | Page and event analytics | When enabled |
| Nango | OAuth for Search Console, Slack, Notion | Connection tokens, held by Nango; our database keeps only the Nango connection id | When enabled |
| Slack and Notion | Recipients of your weekly report (only if you connect them) | Score, fix list, wrong prices and paste-ready snippets for your brand, posted to the one channel or page you chose | Beta |
| Linear | Delivery of your fix list as Linear issues (only if you connect it) | Fix list text and evidence for your brand (engine names, claims, cited URLs, paste-ready snippets); we hold your Linear API key encrypted | When enabled |
| Stripe | Subscription billing | Payment methods and invoices (never touch our servers) | When enabled |
Pre-filled questionnaire
# RulerIQ: vendor security datasheet (draft, 2026-11) - **Product:** RulerIQ (https://ruleriq.com): AI search visibility measurement and fix list - **Scope:** Reads public web pages only. RulerIQ holds no security certifications of its own. ## 1. Data scope - **What do we store?** Scan results (public-page facts and engine answers), the email and domain you submit, IP addresses for rate limiting, and (when enabled) analytics events. Card data is handled by Stripe when billing is enabled. See https://ruleriq.com/privacy. - **Your own API keys:** for one-off checks, a key you supply is used for that request only. For own-key plans (beta), provider keys you save are stored encrypted at rest with AES-256-GCM on the same storage as connection secrets below, shown only as the last 4 characters after saving, used only to run your checks, and deletable any time in Settings. - **Connection secrets (Linear API key and own-key provider keys):** stored encrypted at rest with AES-256-GCM at the application level, using a key held as a Fly secret (never in the database or the code). Shown only as the last 4 characters after saving, never returned by our API, used only to deliver to that service, and deletable any time. They are decrypted only in the background worker that runs delivery and, for one request, in the web app when you press Test or pick a team. Linear delivery is built but paid plans are not open yet. - **MCP access keys:** the personal access keys that let Claude or Cursor read your own tracked brands are issued by us, so they are stored as a SHA-256 hash only (never the key itself), shown once when created, limited to your own tracked brands, rate-limited, and revocable any time. A revoked key stops working immediately. The key is accepted only in the Authorization header, never in a URL. Keys read by default; a key you create with write access can also mark fixes done or undone on your own tracked brands — nothing else. - **Slack and Notion connections (Nango OAuth):** we do not store Slack or Notion tokens. Nango holds them; our database keeps only the Nango connection id and display details (channel name, workspace name, chosen Notion page). Slack access is one channel you pick at install; Notion access is the pages you share. Disconnecting in RulerIQ deletes the connection at Nango, and you can also revoke RulerIQ in Slack or Notion at any time. Beta, for tracked brands. - **What pages do we fetch?** Public marketing URLs on the domain you submit: the homepage, /pricing, /robots.txt, /llms.txt and sitemap files, honoring the site's robots.txt (RulerIQBot rules, else the general * rules). See https://ruleriq.com/bot. - **Do we sit in your CDN/DNS path?** No. - **Do we ingest visitor logs, cookies or visitor IPs from your site?** No. - **Do we need write access to your CMS or codebase?** No. Paste-ready fixes are inspectable text you paste yourself. ## 2. AI providers - **How are engines queried?** Through commercial developer APIs (OpenAI, Anthropic, Google and Perplexity; DeepSeek is not used at launch), not scraped consumer apps. - **What is sent?** Synthetic buyer prompts built from public category and brand names. - **Training:** Check each provider's current API terms; we do not control their retention or training policies. ## 3. Infrastructure and subprocessors - **Hosting:** Fly.io (web app and background workers) - **Database:** Fly Postgres - **LLM processors:** OpenAI, Anthropic, Google, Perplexity (DeepSeek: not used at launch) - **Bot protection:** Cloudflare Turnstile - **Accounts:** Clerk (when enabled) - **Transactional email:** Resend (when enabled) - **Product analytics:** PostHog (when enabled) - **Integrations OAuth:** Nango, for Google Search Console, Slack and Notion (when enabled) - **Slack and Notion:** receive your weekly report content (Score, fix list, wrong prices, paste-ready snippets), only if you connect them (when enabled) - **Linear:** receives the fix list text and evidence for your brand, only if you connect Linear (when enabled) - **Payments:** Stripe (when enabled). RulerIQ does not handle card numbers. - **Network hardening:** SSRF-safe fetcher with DNS pinning that blocks private, loopback, link-local, CGNAT and IPv6 private ranges. ## 4. Certifications - RulerIQ holds no SOC 2, ISO 27001 or similar certification. Questions: support@ruleriq.com
Security questions or reports: support@ruleriq.com. See also Privacy.