Frequently asked questions

Gateways move traffic and dashboards report after the fact. Halo combines discovery on browsers and devices, gateways for model and tool traffic, a secure agent runtime, policy, budgets, and evidence in one control plane.

A Halo agent can research, summarize, draft, organize information, chat with users, and work through approved connected tools. The exact actions depend on the tools, data, policy, and approvals configured for that agent.

Halo-managed access offers a curated catalog of models. Eligible plans can also bring their own provider key, for example for OpenAI, Anthropic, or Google Gemini, while keeping model activity inside Halo policy, gateway, analytics, and budget controls.

Agents can reach 1,000+ apps available through the Halo connection catalog, including Slack, GitHub, Jira, Notion, HubSpot, and Zendesk. Available actions vary by app, and each connection is scoped to the job.

Halo Browser for Chrome and Edge, plus Safari on macOS and Halo Endpoint for macOS and Windows identify AI websites, extensions, desktop apps, developer tools, local runtimes, and add-ins on enrolled devices, then organize them into an inventory with source and confidence.

Standard collection focuses on application, extension, runtime, model, and configuration signals. Prompts, responses, documents, uploaded files, clipboard content, and credentials are not part of standard collection, and expanded browser permissions require your approval.

Halo Browser supports Chrome and Edge, plus Safari on macOS. Halo Endpoint supports macOS and Windows. Rollout is guided so you can start with a pilot group.

Yes. The AI inventory exports as an AI bill of materials in CycloneDX, JSON, or CSV for security, procurement, and audit teams.

Halo gateways apply identity, policy, credentials, approvals, usage, and evidence to approved agent traffic. Each agent receives scoped capabilities for its assigned work, and credentials stay at the gateway.

Yes. Policies can require approval for higher-impact actions, so the agent pauses until an approver decides. The request, the decision, and the outcome are recorded.

Policy evaluates the request at runtime. Halo can allow it, deny it, raise an alert, or pause it for a person to approve, and it records the decision with the reason so reviewers can see what happened.

Operators can quarantine an agent to stop its activity, review the evidence, adjust its access or policy, and release it when ready. Default-deny egress limits where an agent can send data in the meantime.

Policies are versioned and written as code. You can run a policy in observe mode to see what it would do, then switch to enforce mode when you are confident in the results.

Halo records policy decisions, approvals, denials, costs, and configuration changes. Audit logs and gateway decisions export as CSV or JSON for reviewers and other tools.

No software makes an organization compliant on its own. Halo records the inventory, decisions, and approvals that your governance program can draw on when you work toward frameworks such as the EU AI Act, NIST AI RMF, and ISO/IEC 42001.

You can sign up and launch a first secure agent through the self-serve quick deploy path. Browser and endpoint discovery rolls out with guided onboarding, usually starting in observe mode before teams turn on warnings or enforcement.

Agent plans start at $99 per month for one secure agent, with Professional and Team plans for more capacity. Enterprise agreements start at $39,999 per year. Discovery is priced with guided onboarding.

Yes. Our team can help identify valuable use cases, create and connect agents, establish practical governance, and optimize your operating model as adoption grows.