Skip to main content

Cinchy

Cinchy sells PeriMind, a control plane that sits between AI agents and a company's systems and checks every action they take. Each request from tools such as ChatGPT, Claude or custom agents is authenticated, authorized, rate-limited, policy-checked, inspected and logged before it reaches an API, application or database. The company also sells a Data Collaboration Platform for connecting enterprise data.

Traction 10/10
Toronto, Canada · HQ
Founded 2017415K+ followers
  • Cybersecurity
  • AI Agents
  • Enterprise Software
  • Software Only
Updated yesterday

Funding

Raised to date

$24.2MRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.

Funding rounds are not available yet.

Founders

1 founder

Paul Patterson

CEO

Product

Problem

AI agents now update databases, call APIs and run workflows inside large companies, and many of those agents operate without security oversight or logging. Traditional gateways route traffic but do not check what an AI action actually does.

Solution

Cinchy's PeriMind is a control plane that sits between AI interfaces and enterprise systems. Every AI action passes through a governed pipeline with six stages: the agent's identity is verified, its permissions are checked, quotas are enforced, policies written in OPA/Rego are evaluated, the payload is inspected, and the action is logged. A central registry catalogues every AI action endpoint, such as MCP servers, skills, APIs and command-line tools, along with its capabilities and owner. Each agent gets its own identity and scoped credentials instead of shared API keys. The audit log is hash-chained so records cannot be altered after the fact, and the system also captures the AI's reasoning chain alongside each action.

Target Audience

Enterprises running many AI agents across large numbers of internal systems, including financial institutions.

Features

  • See every AI connection and endpoint in one registry, with an owner for each.
  • Write fine-grained rules in Rego to control which agents can call which tools and with what parameters.
  • Set per-agent, per-tool rate limits and quotas to stop runaway agents.
  • Keep tamper-proof, hash-chained logs of every AI action for compliance and incident response.
  • Capture the AI's reasoning behind each action, not only the action itself.
  • Give each agent a unique identity with scoped credentials instead of shared API keys.
This profile is AI-generated and may contain inaccuracies.