Vox for security research

Vox is a terminal agent for security researchers and reverse engineers working in environments where privacy matters. It uses your self-hosted OpenAI-compatible inference server to help investigate code, firmware, and binaries with the tools available in your research environment.

Coordinate an investigation

Give Vox a research question and the material to examine. It can assign parts of the investigation to agents, follow their progress, and bring their findings back for review.

An agent can inventory firmware while another examines selected binaries. A reviewer can check the evidence and identify what needs further work. Each assignment can continue through follow-up questions without starting from scratch.

Work in prepared research environments

Vox can place agents in research workspaces equipped for their assignments, including environments with reverse-engineering tools. Tailor the tools and scope of work to the investigation, then exchange evidence and reports between agents as the research develops.

You can inspect the work from the terminal and dashboard, examine the supporting files, and return to saved investigations later.

Keep the research under your control

Inference runs on infrastructure you operate. Vox has no analytics or telemetry. You decide what material to provide, what actions to permit, and which connections the research environment allows.

Explore private inference, agent orchestration, and research environments, or see how they come together in a research workshop.