Archestra

Archestra Inc. is a London-based technology startup that provides an open-source platform for the secure deployment and integration of AI agents within enterprise environments. Founded in 2024 by CEO Matvey Kukuy and CTO Ildar Iskhakov, the company emerged from stealth mode in August 2025 after securing $3.3 million in a pre-seed funding round led by Concept Ventures. The founding team, which also includes Founding Engineer Joey Orlando, relocated from Israel, Singapore, and Canada to establish the company in the UK.

The founders, Kukuy and Iskhakov, are childhood friends and serial entrepreneurs with a history of successful ventures. They previously founded the incident management platform Amixr, which was acquired by Grafana in 2021. Subsequently, Kukuy co-founded the AIOps platform Keep, which was acquired by Elastic. This background in building and scaling infrastructure and open-source tools provided them with firsthand experience of the security challenges involved in integrating systems, which directly informed Archestra's mission.

Archestra's core product is an enterprise-grade platform that functions as an "MCP orchestrator," designed to safely connect AI agents, such as Large Language Models (LLMs), with sensitive company data and internal systems like HR platforms, email, and Slack. The platform addresses security and compliance risks by introducing layers of security, orchestration, guardrails, and dynamic permission controls. This allows organizations to set clear boundaries for AI agent behavior, preventing issues like unauthorized data access or accidental file deletion. The solution is built upon the Model Context Protocol (MCP), a framework introduced by Anthropic that facilitates the connection of LLMs with various data sources, making it enterprise-ready.

The platform is designed to be accessible to both technical and non-technical users, enabling employees across different departments, such as HR or finance, to integrate AI agents into their workflows without compromising security. Key features include a secure runtime environment with sandboxing, advanced prompt injection prevention, and a local-first architecture for privacy. Archestra is framework-agnostic, compatible with tools like LangChain and LlamaIndex, and can be deployed on-premises in a Kubernetes cluster for maximum data control. The company operates within the integration-platform-as-a-service (iPaaS) market and provides its open-source solution to build community trust and drive adoption, with plans to offer enterprise-grade features.

Keywords: AI agent security, Model Context Protocol, MCP orchestrator, enterprise AI integration, open-source security, AI guardrails, data governance, secure AI deployment, LLM security, dynamic permissions, context management, AI workflow automation, integration-platform-as-a-service, iPaaS, AI safety, prompt injection prevention, enterprise software, AI infrastructure, access control, compliance, data protection, agentic gateway, autonomous agents, AI runtime