What can Nexus do in my browser?

Nexus can read normal web pages, search public sources, navigate, click, type, scroll, take screenshots, and coordinate work across tabs. Chrome internal pages, the Web Store, and other protected extension pages remain off limits.

Which AI providers does it support?

Anthropic, OpenAI, xAI Grok, OpenRouter, DeepSeek, and local Ollama. You bring your own key, endpoint, and model; a connection test verifies the setup before your first task.

How does Nexus keep automated actions safe?

Choose autonomous, confirm-risky, or confirm-all mode, add sensitive domains to the blocklist, and review the live action trace. Risky actions such as form submission, destructive clicks, and tab closing can require explicit approval.

Does Nexus send my browsing history to Selfhood?

No. Nexus has no Selfhood backend or telemetry. Keys, history, memories, and task records stay in Chrome storage; only content needed for a task is sent directly to the model provider you selected.

NexusProductivity

Nexus — Give your browser a task. Watch it get done.

An agentic AI workspace inside Chrome. Ask in plain language and Nexus can research, navigate, click, type, and coordinate work across tabs — with visible steps, safety controls, and your choice of AI provider.

V1.0.0ProductivityAI BYOKLocal-first

Nexus — coming soon

Interactive · no install needed

Feel it before you add it

Feature tour

What Nexus actually does

Command your browser

Describe the outcome, not the clicks

Give Nexus a task in everyday language and it works through the browser for you — opening and switching tabs, navigating, clicking, typing, and extracting information from the pages you use.

Natural-language tasks in a persistent side panel

Real browser input through the Chrome Debugger Protocol

Multi-tab navigation, reading, and action orchestration

Verifiable actions

See every step while it happens

Nexus keeps its work inspectable. A live step trace shows each search, page read, navigation, and browser action, while an on-page indicator makes it clear whenever the agent is controlling a tab.

Expandable trace for every action and result

Cited headless research for public information

Visible on-page status and target highlighting

Mid-task steering

Redirect the work without starting over

Add guidance by typing or speaking while Nexus is running. Your note lands in the next reasoning step, so the agent can adjust course while preserving the task, context, and progress it already has.

Text or voice guidance during an active run

Retry failed or stopped work in place

Saved task history for quick reruns

Your model, your keys

Choose the intelligence behind the agent

Connect Anthropic, OpenAI, xAI, OpenRouter, DeepSeek, or a local Ollama model. Nexus uses one provider-neutral agent loop, while your credentials remain in Chrome storage on your device.

Six provider routes, including local Ollama

Bring your own key and choose any supported model

Live token and cost estimates while tasks run

Choose the intelligence behind the agent

Safety & audit

Automation with deliberate guardrails

Choose how much autonomy Nexus has, require approval for risky or every action, and block sensitive domains outright. When the work is done, export the full task record as a readable Markdown audit log.

Autonomous, confirm-risky, and confirm-all modes

Domain blocklist and prompt-injection defenses

Exportable prompts, actions, results, and token totals

Command your browser