Git Security Alternative

Pretense vs GitHub Secret Scanning

GitHub Secret Scanning protects your git repository. It does nothing to protect code you send to Copilot, Cursor, or Claude Code. Pretense fills that gap with pre-send mutation.

9 of 10features won
$29per seat per month
30 secsetup time
Local-firstnothing leaves your machine
vs

GitHub Secret Scanning

GitHub Secret Scanning scans git commits and pull requests for secrets like API keys and tokens. It alerts repository owners when exposed credentials are found. It is a post-commit, git-layer tool and has no mechanism to protect what developers actively send to AI coding assistants.

Pretense

Pretense operates at the AI API layer, not the git layer. It intercepts every request your AI coding tool sends to LLM providers and mutates proprietary identifiers before transmission. This protects both code content and secrets that never touch git.

Feature Comparison

Side-by-side view of how Pretense and GitHub Secret Scanning compare on the capabilities that matter most to security teams.

Feature
Pretense
GitHub Secret Scanning
Where it operates
AI API layer (pre-send)
Git layer (post-commit)
Protects code sent to AI tools
Yes, core use case
No, does not apply
When it acts
Before LLM receives code
After code is committed to git
Covers code never committed to git
Yes
No
Mutation of proprietary identifiers
Yes, full identifier mutation
Not applicable
Works with Copilot, Cursor, Claude Code
Yes, native integration
No coverage
SOC2 AI usage audit log
Yes, built-in
No (git audit only)
Compliance report export
PDF and JSON export
Not available for AI usage
Price per developer
$29 per seat per month
Free (included in GitHub Advanced Security at $49 per seat)
Local-first
Yes
No (GitHub cloud)

Why Teams Switch from GitHub Secret Scanning

Two different threat surfaces

GitHub Secret Scanning protects committed code in repositories. Pretense protects code in flight to AI APIs. These are different threat surfaces. A developer can paste proprietary code into a Copilot prompt, have it sent to Microsoft servers, and GitHub Secret Scanning never fires because nothing was committed. Pretense catches that.

AI tools see code that never enters git

Developers frequently ask AI tools about code that exists only in their working directory. Untracked files, scratch work, and prototype code all get sent to AI providers without ever touching a git commit. GitHub Secret Scanning has zero visibility here. Pretense protects all of it.

From reactive to proactive

Secret scanning is a reactive control. It finds what leaked after the fact. Pretense is a proactive control. It prevents exposure before it occurs. For proprietary algorithms, business logic, and trade secrets, proactive is the only acceptable posture.

Frequently Asked Questions

Should I use both GitHub Secret Scanning and Pretense?

Yes, they are complementary. GitHub Secret Scanning protects your git repositories. Pretense protects what developers send to AI tools. Both controls are needed for comprehensive coverage.

Does Pretense replace push protection in GitHub?

No. GitHub push protection prevents secrets from being committed to repositories. Pretense prevents secrets and proprietary code from being sent to LLM APIs. These protect different parts of your attack surface.

Can Pretense detect secrets in AI prompts?

Yes. Pretense runs 30+ secret patterns against every prompt before transmission. API keys, tokens, connection strings, private keys, and PII are blocked at the edge with a clear error message.

Ready to switch from GitHub Secret Scanning?

Install Pretense in 30 seconds. One environment variable. No code changes. Protect every AI tool request from day one.

No credit card required. Free tier available. Local-first, nothing leaves your machine.

Ask me anything