AI security for enterprise engineering teams

Your developers use AI.
Your IP stays private.

Pretense is an AI firewall that mutates proprietary code before it reaches any LLM. Real identifiers never leave your network. Deploys in 30 seconds. Works with every AI tool your team already uses.

Open source600+ testsSOC2 in progressNo cloud dependency

The threat is real

Every day, proprietary code leaks to LLM providers through AI coding tools. Traditional DLP can't stop it — it doesn't understand code.

0

Samsung IP exposure from a single ChatGPT incident

0

of developers share proprietary code with AI tools weekly

0

of Copilot suggestions contain private data from other repos

Works with every AI tool your team uses

ClaudeCopilotCursorCodexOpenAIAnthropicGeminiOllama

How it works

Three commands. Full protection.

1

Install

curl -fsSL pretense.ai/install.sh | sh

macOS, Linux, WSL. No dependencies.

2

Scan & configure

pretense init

Scans your repo. Creates mutation config.

3

Protected

pretense start

Local proxy. All AI traffic mutated.

Point your AI tools at localhost:9339 and every identifier is mutated before it leaves your machine. Responses are automatically reversed.

Why Pretense

The only AI firewall that mutates code

Traditional DLP blocks or redacts. Pretense mutates — so the LLM still gets working code, but your identifiers are synthetic.

Code mutation, not redaction

Every identifier — functions, variables, classes — is deterministically mutated to synthetic tokens. The LLM gets working code it can reason about. Your IP stays private.

Secrets blocked at the edge

API keys, tokens, passwords, and PII are detected and blocked before they leave your machine. 30+ patterns covering every major secret format.

Enterprise-grade compliance

SOC2, HIPAA, GDPR, and PCI-DSS ready. Full audit trail of every mutation. Immutable logs for your security team and regulators.

Enterprise AI security, compared

FeaturePretenseTraditional DLPNo Protection
Code mutation (preserves AI context)
Deploys in 30 seconds
Works offline (local-first)
Multi-provider (Claude + GPT + Gemini)
SOC2/HIPAA audit trail
Starting at $0/month$5K+/mo

Enterprise DLP solutions start at $5K/month. Pretense is free for teams up to 3 seats.

SOC2 In ProgressHIPAA DesignedGDPR DesignedPCI-DSS Compatible

Works with every major LLM and IDE

C
Claude
G
ChatGPT
Gm
Gemini
VS
VS Code
Cu
Cursor
GH
GitHub

Deploy AI security in 30 seconds

Local-first. Open source. No credit card required.
Free for teams up to 3 seats.