Bay
by Bay AI, Inc.
AI agent security platform that discovers every agent and MCP server and blocks risky agent actions at runtime
Bay is an AI agent security platform for enterprise security teams. It finds the AI agents, MCP servers, plugins and skills running on laptops, browsers and cloud workspaces, then decides in milliseconds whether each agent action is allowed, blocked or sent to a human for approval.
Bay is an Israeli security startup that came out of stealth on November 10, 2026 with $50 million: a $37 million Series A led by Battery Ventures, with Micron Ventures, Microsoft's M12 and Wiz CEO Assaf Rappaport, plus a previously unannounced $13 million seed led by YL Ventures and Striker Venture Partners. Its co-founders are CEO Yuval Zacharia, previously head of research and AI at Hunters after leading a department in Israel's Unit 8200; CPO Aner Morag, formerly VP of product at Noname Security; and VP of R&D Boris Danilovich, who led engineering teams at CrowdStrike. The product is split into Endpoint, Cloud Workspace and Browser modules. Bay discovers every agent, MCP server, plugin and skill in use and shows what each can reach and who uses it. It flags posture risks such as toxic tool combinations, exposed credentials and disabled sandboxes. At runtime it evaluates each proposed action against the whole session and returns allow, ask or deny. It learns per-company agent behaviour, so security teams do not have to write rules for every new tool. Deployment is agentless through existing EDR or MDM infrastructure, with no kernel driver or network proxy, and Bay claims visibility within five minutes. It holds SOC 2 and ISO 27001. Customer logos include Micron, Peloton, ClickUp, Wix and The Motley Fool. Competitors include Zenity, Straiker, NeuralTrust and Asymptote Labs.
A CISO or security engineering lead whose developers and business users are already running coding agents, MCP servers and browser AI tools that nobody has inventoried.
A full inventory of agents and MCP servers in use, plus runtime blocking of risky agent actions without deploying a new endpoint agent.
At a Glance
- Category
- Governance & Security
- Pricing
- Contact for pricing
- Target Market
- CISOs, Security Engineering, CIOs, IT Leaders
- Deployment
- Cloud-first
- Headquarters
- Tel Aviv, Israel
- Team Size
- 11-50
- Customers
- Dozens of enterprise customers (company claim, Nov 2026)
Key Features
- ✓Shadow AI discovery
Inventories every AI agent, MCP server, plugin and skill across the organisation, showing what each can access and who uses it.
- ✓AI security posture
Checks agent configuration for toxic tool combinations, exposed credentials and disabled sandboxes, with automated remediation of the findings.
- ✓Agent runtime enforcement
Evaluates each proposed agent action against the full session and allows, blocks or routes it to a human in milliseconds.
- ✓Adaptive guardrails
Learns each company's agent behaviour and applies different policies per user group, such as developers versus finance, without hand-written rules.
- ✓Agent investigation and forensics
Keeps an audit trail of agent actions so security teams can reconstruct what an agent did during an incident.
- ✓Agentless deployment
Rolls out through existing EDR or MDM tooling with no kernel driver or proxy, which Bay says gives visibility within five minutes.
Use Cases
- •Inventory of AI tools in use
A security team switches Bay on and gets a map of every agent and MCP server running across laptops and browsers within minutes.
- •Blocking destructive agent actions
When a coding agent tries a risky command or data access, Bay blocks it or asks the user to approve before it executes.
- •Scaling sanctioned agent use
With guardrails in place, IT can approve broader agent rollouts instead of banning tools, which Bay says customers used to expand agent use.
- •AI governance reporting
Compliance teams pull governance reports and agent activity logs to show auditors how autonomous AI use is controlled across the company.
Ideal For
Best For
- ✓Shadow AI discovery across employee laptops, browsers and SaaS workspaces
- ✓Governing coding agents and MCP servers that developers install without security review
- ✓Runtime enforcement that blocks or escalates risky agent actions instead of only alerting
- ✓Audit trails and forensics of agent activity for compliance and incident investigations
Not Ideal For
- ✗Teams that need published pricing or a self-serve trial; Bay sells by demo only and lists no prices
- ✗Buyers who want a long track record and peer reviews, since the company left stealth in November 2026 and has no G2 or Hacker News presence yet
- ✗Organisations without an existing EDR or MDM footprint, which Bay relies on for its agentless deployment
Deployment
Market & Ratings
Dozens of enterprise customers (company claim, Nov 2026)
Market Analysis
Pros
- ✓Covers endpoints, browsers and cloud workspaces in one product rather than a single surface
- ✓Agentless deployment lowers rollout friction for security teams already running EDR or MDM
- ✓SOC 2 and ISO 27001 certified at launch
- ✓Founders bring security product experience from Hunters, Noname Security, CrowdStrike and Unit 8200
Cons
- ✗Its headline claim, that Fortune 100 companies tripled autonomous agent use, names no customers or method, and TNW could not verify it
- ✗Micron is both a named customer and an investor through Micron Ventures
- ✗Crowded category with better-funded rivals such as Zenity ($125M Series C) and Straiker ($64M Series A)
- ✗No public pricing, documentation, supported-agent list or user reviews yet
Pricing
Bay AI Agent Security Platform
Contact for pricing
- ✓Endpoint, Cloud Workspace and Browser modules
- ✓Shadow AI discovery and posture
- ✓Runtime enforcement
- ✓Investigation and compliance reporting
Bay publishes no list pricing as of November 12, 2026; the site offers only a demo request. The metering unit (per device, per user or per agent) and which modules are bundled are not disclosed, so buyers should ask how Endpoint, Cloud Workspace and Browser are priced separately.
Security & Compliance
Connect
Sources
This page was written from 5 sources, 2 on domains other than bay.io.
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Larridin
AI value measurement platform that tracks enterprise AI spend, adoption and output across human and agent work
Cyara Agentic CX Trust Platform
Independent testing, monitoring and observability for AI agents in voice and digital customer service
Sysdig AI Defense
Runtime security and governance that inventories AI agents and blocks risky actions before they complete
Sentra
AI data security platform that maps which people and agents can reach sensitive data and removes excess access