pyobfus — Python Obfuscator

What is pyobfus — Python Obfuscator?

pyobfus is an AST-based Python obfuscator MCP server that lets AI agents (Claude Code, Cursor, Windsurf, Zed) protect Python source code end-to-end from a conversation. It serves as an open-source, no-phone-home alternative to PyArmor.

How to use pyobfus — Python Obfuscator?

Install via uvx pyobfus-mcp or pip install pyobfus-mcp. No API key is required. Once installed, an AI agent can call any of the 8 built-in tools to obfuscate projects, scan risks, generate configs, or reverse obfuscated tracebacks.

Key features of pyobfus — Python Obfuscator

• AST-based obfuscation with no phone-home.
• 8 tools including protect_project, check_obfuscation_risks, and unmap_stack_trace.
• All tools return structured JSON with ai_hint and next_tool for deterministic chaining.
• Preflight risk scanning for eval/exec, dynamic attributes, and framework reflection.
• Auto‑generates pyobfus.yaml based on detected framework.
• Open‑source alternative to PyArmor with no API key required.

Use cases of pyobfus — Python Obfuscator

• Obfuscate a complete Python project and verify it still runs in one MCP call.
• Scan a project for obfuscation risks before applying protection.
• Auto‑generate configuration files tuned to the project’s framework.
• Debug deobfuscated stack traces from obfuscated code during development.

FAQ from pyobfus — Python Obfuscator

How does pyobfus compare to PyArmor?

pyobfus is an open-source, AST-based obfuscator that does not phone home, making it a transparent alternative to PyArmor.

Does pyobfus require an API key?

No, pyobfus does not require any API key.

How do I install pyobfus?

Install using uvx pyobfus-mcp or pip install pyobfus-mcp.

What tools does pyobfus provide?

It provides 8 tools: protect_project, check_obfuscation_risks, generate_pyobfus_config, unmap_stack_trace, list_presets, explain_preset, recommend_tier, and start_pro_trial.

What format do the tools return?

All tools return structured JSON that includes an ai_hint and a next_tool field for deterministic chaining of operations.