Clawvisor
An open-core authorization gateway that scopes and audits what AI agents are allowed to do before granting them live credentials.
Visit Website ↗ + Add to Compare Claim This CompanyOverview
Clawvisor is an authorization layer and control plane that sits between AI agents and the applications they act on, such as Gmail, Slack, Google Drive, GitHub, and Stripe. Rather than handing agents standing credentials, a user approves a task — defined as a purpose plus the specific tools it requires — and Clawvisor evaluates every subsequent request against that approved scope before injecting short-lived, scoped credentials and logging the action. Requests that fall outside the approved scope are routed to a human approval queue instead of executing automatically.
The San Francisco company is a solo-founder effort by Eric Levine, who previously co-founded identity-verification company Berbix (acquired by Socure in 2023) and led trust-and-safety engineering at Airbnb. Clawvisor is open-core, with its gateway published under the Elastic License 2.0 on GitHub, and ships as a self-hostable Go daemon with a CLI, web dashboard, TUI, and an OAuth 2.1 MCP server, including an official Claude Code plugin. It is part of Y Combinator’s Spring 2026 batch.
Innovation Matrix Assessment
A solo founder has already shipped a fairly complete open-core stack (daemon, CLI, dashboard, MCP server, Claude Code plugin) within the YC S26 cycle.
Directly addresses a concrete, fast-emerging CISO concern — AI agents holding broad standing credentials — with a task-scoped, short-lived-credential model.
Extremely early stage: one employee, no disclosed funding amount, and no named enterprise customers found.
An agent-native authorization layer is a genuinely different approach from retrofitting legacy IAM for agents, though the AI-agent-security category is crowded with similarly positioned new entrants.
The open-source codebase is externally inspectable, which is a plus, but there is no independent audit, named customer, or incident evidence yet.
As agentic AI deployments scale across enterprises, scoped authorization and credential isolation for agents is likely to become a core control, not a niche one.
Why CISOs Should Care
Gives security teams a way to let AI agents act on real systems (email, code, payments) without granting them standing, unscoped credentials.
What Makes It Different
Agents never hold underlying credentials directly; every action is checked against a user-approved task scope and logged, with drift routed to human review.
The Matrix Verdict
57/100 — INCREMENTAL INNOVATOR
A very early but conceptually sound entrant in the fast-growing AI agent security space; an Incremental Innovator today, with its trajectory entirely dependent on customer traction it has not yet demonstrated.
Editorial Note: Claims vs. Verified Findings
Product and architecture details come from the company's own site and YC listing; no independent customer or security-audit evidence was found.
Sources
Alternatives to Clawvisor
Quilr
CISO ReviewedEarly-stage agentic AI security startup building a 'Service-as-Software' platform to guard against human-related breaches and secure AI agent…
Adaptive Security
AI-driven platform that simulates deepfake, voice, and multichannel social-engineering attacks to train and test organizations against next-generation phishing.
Zenity
Governance and security platform for AI agents and low-code/no-code development, securing agent identity, permissions and behavior across the…
Tenzai
An agentic AI penetration testing startup building autonomous 'AI hackers' to find and validate exploitable vulnerabilities at a…
Gray Swan AI
Carnegie Mellon-linked startup building automated red-teaming and runtime guardrails that frontier AI labs use to stress-test their models…
Charm Security
Agentic AI workforce that investigates and intervenes on scams and fraud in real time, reading manipulation and intent…