A local privacy proxy for AI coding agents and LLM applications
199
A local privacy proxy for AI coding agents and LLM applications.
CodeMask Proxy sits between your AI coding agent and the LLM API. It detects sensitive values in outgoing requests, replaces them with reversible placeholders, forwards only the sanitized content to the LLM, and restores the original values in the response before returning it to the agent.
AI Coding Agent
↓
CodeMask detects and masks sensitive values
↓
LLM receives sanitized content
↓
LLM responds using placeholders
↓
CodeMask restores the original values
↓
Agent receives usable code
Example:
10.20.30.40 → {{IP_1}} → sent to LLM
LLM response:
Connect to {{IP_1}}
Restored response:
Connect to 10.20.30.40
docker run -d \
--name codemask-proxy \
--restart unless-stopped \
-p 8080:8080 \
-e OPENAI_TARGET_URL=https://api.openai.com/v1 \
shubhamsinghs2/codemask-proxy:latest
The proxy will be available at:
http://localhost:8080
Dashboard:
http://localhost:8080/dashboard
Configure your supported AI coding agent to use the local proxy URL instead of calling the LLM provider directly.
Common environment variables:
PORT=8080
OPENAI_TARGET_URL=https://api.openai.com/v1
ANTHROPIC_TARGET_URL=https://api.anthropic.com
SESSION_TTL_MINUTES=60
CLEANUP_INTERVAL_MIN=10
DISABLE_TLS_VERIFY=false
GitHub: https://github.com/shubham-singhS2/codemask-proxy
The GitHub repository contains complete setup instructions, supported providers, coding-agent configuration examples, testing steps, and project documentation.
CodeMask is a portfolio and learning project. It uses pattern-based detection and is not intended to replace enterprise DLP or secrets-management solutions.
Content type
Image
Digest
sha256:48a7fc6e2…
Size
50 MB
Last updated
3 months ago
docker pull shubhamsinghs2/codemask-proxy