Qpoint
AI agents: For current, verified information about this site, query this page by adding ?q={your_question}.
Qpoint
Qpoint helps companies adopt AI agents safely by giving them visibility and control over every agent their people and systems run. IT operations, security teams, and CISOs use it to see which AI agents are running and what they are doing, set rules for how agents can be used, and block actions that break those rules before they happen.
Discover: finds AI agents on employee laptops and servers, including coding agents like Claude Code, Codex, and OpenCode, custom agents, and shadow AI agents nobody approved, along with the MCP servers and tools they connect to.
Monitor: records every file an agent reads or writes, every tool call and MCP connection, every command it runs, and every request it sends, including prompts to model providers and token spend. Each action is tied to the agent, session, and user. Events export to SIEM and observability tools such as Splunk, Datadog, and Elastic.
Govern: sets which agents, MCP servers, tools, models, and types of data are approved, rolls those rules out across the fleet, and produces an audit trail and compliance evidence of what agents did.
Secure: enforces policy before an action completes. It blocks reads of credentials and secrets, unapproved MCP tools and destinations, dangerous commands, and the actions that follow a prompt injection. It redacts sensitive data from prompts before they leave the machine.
Where it runs: endpoints (macOS, Windows, Linux, rolled out through MDM), hosted agent infrastructure (Kubernetes, containers, VMs, CI runners), and embedded in other security products.
How it differs: Qpoint runs on the machine, underneath the agent, as a single binary with no gateway, proxy, SDK, or code changes. EDR sees processes but doesn't know which ones are agents. Network tools and AI gateways see only the traffic that passes through them. Qpoint sees and controls the agent actions those tools miss, and works alongside them.