Security & Trust

GSCPilot is built on the principle of minimal access. We only request the permissions we need, encrypt everything at rest, deliver changes as reviewable pull requests, and run on trusted, audited infrastructure.

01

Authentication & Access Model

GSCPilot uses a GitHub App for repository access, not personal access tokens. The App follows the Principle of Least Privilege: it requests only the permissions needed to read files and create pull requests.

For Google Search Console, we request webmasters.readonly Read-only access. We never modify your Google properties, analytics, or any other Google service.

You control which repositories the GitHub App can access at any time via your GitHub settings.

02

Pull Request Safety Model

By default, every code change GSCPilot generates is delivered as a pull request on a new branch. You review the full diff before merging. For teams that prefer it, a direct push mode is available in site settings, but PR mode is the default.

We do not:

  • Auto-merge pull requests
  • Rewrite git history or force-push
  • Modify files outside the scope of the SEO fix
  • Access repositories you have not explicitly granted

Whether you use PR mode or direct push, every change is scoped to the specific SEO fix and nothing else.

03

Data Encryption

All OAuth tokens (Google, GitHub) are encrypted at rest with AES-256-GCM before storage. HTTPS with TLS 1.2+ is enforced on every endpoint. There is no unencrypted path to our servers.

Secrets and API keys are stored in Vercel's encrypted vault, isolated from application code and never committed to source control. Database connections use encrypted channels with managed credentials.

04

Infrastructure

GSCPilot runs on boring, trusted infrastructure:

  • Vercel: SOC 2 Type II compliant hosting and serverless functions
  • Supabase: SOC 2 Type II, HIPAA eligible database and authentication
  • Vercel AI Gateway: Managed AI routing, no direct API key exposure

No data is stored on self-managed servers. All infrastructure runs on audited cloud platforms with automated security patching.

05

AI Processing & Data Privacy

AI models process your page content to generate SEO recommendations. Only the minimum data needed is sent to AI providers during active scans:

  • Page titles, meta descriptions, headings, and schema markup
  • Public Google Search Console metrics (clicks, impressions, CTR)

We do not:

  • Train AI models on your data
  • Send private source code, credentials, or payment info to AI providers
  • Store AI processing results beyond the generated recommendations
  • Share your data with third parties for marketing or advertising

AI providers: Google (Gemini), Anthropic (Claude), and OpenAI (GPT), all via inference-only API calls through Vercel AI Gateway.

06

Data Retention & Deletion

We retain only what's needed to operate the service:

  • OAuth tokens: encrypted in database, revoked when you disconnect or delete your account
  • GSC snapshots: aggregated performance data, retained while your site is connected
  • Recommendations: stored until acted upon or your account is deleted
  • Logs: server logs retained for 14 days maximum

You can delete your account and all associated data at any time from your dashboard settings. Deletion is immediate and permanent. All sites, recommendations, changes, tokens, and usage data are removed.

Security questions?

We're happy to answer any security or compliance questions. Contact alex@athorp.io