About
ContextFort builds browser and endpoint security tools for security teams managing AI agents. Its Chrome extension detects when agents control browsers, records actions such as pages visited, clicks, and text entry, and lets teams block risky actions and cross-site data leakage.
Market
ContextFort competes in the emerging AI-agent security market, focused on endpoint and runtime visibility and controls for browser and coding agents. It differentiates from traditional EDR by independently monitoring agent activity at the OS or kernel level, recording tamper-resistant process and action histories, and addressing what an agent accessed or executed rather than only asking whether activity was malicious.
ContextFort primarily targets security teams at organizations deploying AI browser and coding agents. Its likely buyers are responsible for endpoint visibility, policy enforcement, secret-leak prevention, and auditability of agent actions.
At a Glance
Problem
ContextFort addresses the security gap created when AI agents operate with the permissions of the engineer or user who launched them. Coding agents such as Cursor or Claude Code may be able to read source code, environment variables, cloud credentials, SSH keys, and other sensitive files, while also making network connections and executing shell commands. The agent does not need to be malicious for this to create risk: it may read secrets as project context and transmit them to an AI provider. The economic pain is the potential cost of credential exposure, data exfiltration, security incidents, and the inability to demonstrate what an AI agent actually did.
The killer use case is giving security teams an independent record and control layer for AI-assisted software development and browser automation. Instead of merely asking whether a process looks malicious, organizations need to know which files an agent accessed, which commands it ran, which sites it visited, and whether it attempted a risky or cross-site action.
Product / Service
ContextFort is positioned as an AI-agent security product spanning endpoint monitoring for coding agents and browser controls for autonomous agents. Its endpoint product monitors activity independently at the operating-system level, using eBPF on Linux, Endpoint Security Framework on macOS, and ETW plus Minifilter on Windows. This produces tamper-resistant telemetry on file access and agent actions rather than relying only on logs reported by the agent itself.
For browser agents, the company describes a Chrome extension that detects when an agent takes control of the browser, records pages visited, clicks, and text entry, and provides controls to prevent risky actions and cross-site data leakage. The benefit is visibility, auditability, and policy enforcement for AI agents without requiring the organization to trust the agent's own account of what happened; the commercial motion appears to be demo-led, with the site inviting prospective customers to book a demo.
Market
ContextFort competes in the emerging cybersecurity market for AI-agent runtime security, with a particular focus on AI coding agents, endpoint activity, and browser agents. Its relevant alternatives include incumbent endpoint-detection platforms such as CrowdStrike, which ContextFort argues can detect malware but do not independently audit AI-agent behavior, as well as purpose-built AI-agent security vendors such as WitnessAI and Silverfort. ContextFort's differentiation is the combination of OS-level, agent-independent telemetry and browser-session controls rather than conventional malware detection alone.
The company is an early-stage, YC-backed startup from the Summer 2025 batch, founded in 2025 and listed as active with two employees in San Francisco. Public materials do not establish a customer count or verified revenue figure; third-party databases give conflicting revenue and funding signals, although one reports $500,000 of funding. The YC profile also retains an earlier construction-drawing-review launch description, suggesting the company may have pivoted from or expanded beyond that initial construction use case, so its current security positioning is clearer than its publicly disclosed commercial traction.
Founders & Leadership
Funding History
Y Combinator
Recent News
Y Combinator’s 2026 security-startup directory lists ContextFort as an active San Francisco company building visibility and controls for AI browser agents. The listing describes its Chrome extension as detecting when an AI agent takes control of the browser.
Startup Intros’ updated company profile describes ContextFort as an AI copilot for reviewing architectural and structural drawings in construction. It says the platform automates drawing review and RFI generation, helping contractors identify dimensions, conflicts, and trade-specific changes.
ContextFort’s product page presents a browser-agent security product that provides visibility into agent sessions and controls for blocking risky actions, pages, and cross-site flows. The page positions the product as protection for autonomous browser agents and says it is backed by Y Combinator.
Y Combinator’s company launch profile introduced ContextFort as an AI copilot for construction drawing review. The product identifies missing or inconsistent dimensions, finds overlay mismatches, checks whether RFIs are already covered by construction documents, and drafts RFIs from specification sheets; YC says it saves contractors 2–3 weeks per project.
Active Roles
0No active roles right now.
Get notified when they postBusiness Model
Public pricing is not disclosed. The company appears to monetize through direct enterprise software sales, including cloud or self-hosted deployments of its AI-agent visibility and control platform.