ReconX

Agent-driven recon

Instead of memorising flags and stitching tools together, you describe what you want. ReconX picks the tool, runs it safely, and turns raw output into a clear, readable report.

It ships three ways, a full web app, a cross-platform desktop app built with Electron, and a pip-installable Python package, so it fits whatever workflow you already have.

Built to explain, not just scan

At its core is an agent loop that maps a request to the correct recon step, executes it, and feeds the results back for interpretation, keeping a human in control of every action.

One toolkit, three surfaces

The React front end streams results as they arrive, while the FastAPI backend orchestrates the tools and the language model that narrates each finding.

Every finding comes with context: what was checked, what it means, and what to do next. ReconX makes recon approachable without hiding the underlying commands.

Human stays in control

The same engine powers the web app, the desktop build, and the reconx-py CLI, so teams can start in the browser and drop to the command line when they need to.

ReconX is built to grow, new tools and checks slot into the agent without rewriting the UI, and the Python core is open for anyone to extend.

Recon without the flags

Under the hood it is a clean React + FastAPI stack: the front end streams tool output live, the backend orchestrates each recon step, and the language model narrates findings. Everything is modular, so new tools and checks slot in without rewriting the UI.

Findings you can act on

ReconX keeps growing: new tools and checks slot into the agent without touching the UI, and the open Python core means anyone can extend it.

Open and extensible

ReconX keeps the human in the loop: every tool call is visible, every finding is explained, and the same engine powers the web app, desktop build, and Python CLI.