Nexus — AI Browser Agent

Nexus is an agentic AI workspace designed to operate directly within your Chrome browser. It allows users to perform complex tasks by simply describing the desired outcome in plain language. Nexus can research, navigate websites, interact with elements by clicking and typing, and coordinate actions across multiple browser tabs.

Key Features:

• Natural Language Interaction: Command your browser using everyday language, eliminating the need for manual clicks and complex instructions.

• Visible Execution: Monitor the agent's actions in real-time through a live step trace and on-page indicators, ensuring transparency and control.

• Mid-Task Steering: Adjust or redirect the agent's workflow mid-execution by providing new guidance via text or voice, preserving context and progress.

• Multi-Provider Support: Integrate with various AI providers, including Anthropic, OpenAI, xAI Grok, OpenRouter, DeepSeek, and local Ollama instances, allowing you to use your preferred AI models.

• Enhanced Safety Controls: Customize automation levels with options for autonomous, confirm-risky, or confirm-all modes, and utilize domain blocklists and prompt-injection defenses for secure operation.

• Audit Exports: Generate comprehensive audit logs of tasks, including prompts, actions, and results, in a readable Markdown format.

• Local-First Privacy: All sensitive data, including API keys, history, and task records, are stored locally within Chrome, with no backend or telemetry sent to Selfhood Studios.

Nexus empowers users to automate repetitive browser tasks efficiently and securely, all from within their familiar browsing environment.