Qpoint

QPoint develops software that provides visibility into and control over service connections and AI agent activity from the endpoint. Its flagship agent-security product, QControl, deploys as a single binary (started with a `qcontrol start` command) on any endpoint or device that runs AI agents. Within seconds it enumerates every agent on the machine regardless of runtime, framework, or language, then captures the agent's actions at the source: file reads and writes, command and process execution, outbound network requests including DNS lookups and HTTP headers/bodies/responses, MCP server routes and tool calls with input parameters and results, sub-agents, skills, and LLM calls. Events are normalized into a single stream and fed into the customer's existing SIEM, SOC, and compliance tooling. Enforcement is handled through plugins that can observe, block, annotate, or rewrite activity on any surface QControl watches — for example refusing reads on paths such as ~/.ssh, .env files, or credential stores by returning ENOENT, tagging tool calls with user, session, and risk-tier attributes, or redacting secrets from an HTTP request body before it leaves the machine. Plugins written against the QControl SDK are intended to apply across agent harnesses including Claude, Codex, OpenCode, and OpenClaw.

The company's broader product line addresses service-connection visibility. QTap uses lightweight eBPF sensors on Linux to observe requests and responses before TLS encryption occurs, providing payload and header visibility without proxies, certificate management, or application code changes, and attributing each connection to the specific originating process, container, or pod. QPlane targets compliance evidence, generating proof of data residency for audits. A browser-based DevTools product offers live server-side HTTP/S traffic inspection for production debugging. Company materials position the platform for discovering undocumented internal (east-west) and third-party (north-south) dependencies, monitoring SLA compliance and error rates, tracking API usage and costs, detecting PII and secrets in traffic using machine-learning classification, and supporting zero-trust service-to-service policies.

QPoint frames its positioning against adjacent security layers: identity systems know who authenticated but not what the agent did afterward; network perimeter tools see encrypted flows and destinations but not intent; EDR sees process lineage and syscalls but does not recognize the process as an AI agent; and AI gateways only see what crosses the gateway, missing local tool calls, file reads, and spawned subprocesses.

Founding story

Qpoint was founded to address the visibility and control gaps created by the shift from self-contained monolithic architectures to applications that depend on a wide range of external APIs and cloud services, which introduced unpredictable outages, security risks, and complex troubleshooting for operators.

Business model

QPoint sells enterprise software products (QControl, QTap, QPlane, DevTools) deployed on customer endpoints and infrastructure, with a published pricing page, demo requests, and customer support offered to buyers.

Paid commercial software; pricing is published on the company's pricing page and support is provided to customers.

Traction

The company markets multiple shipping products with public documentation, release notes, and deployment guides, and states that QControl installs and begins surfacing agents in under a minute.

Latest developments

QPoint's current positioning centers on QControl, an agent security and governance product that discovers, observes, and enforces policy on AI agents from a single endpoint binary, with a plugin SDK designed to work across agent harnesses such as Claude, Codex, OpenCode, and OpenClaw.