# Agent Enclosure

> Secure sandboxed execution environment for AI agents in production.

Agent Enclosure gives engineering teams an auditable perimeter around each AI agent — scoped permissions, network policies, and automatic kill-switches — so they can deploy capable agents without betting the company on them.

## What is Agent Enclosure?

Agent Enclosure is a security infrastructure product for AI agent deployment. It wraps AI agents (LangChain, AutoGen, CrewAI, custom) with an isolated execution environment that enforces:

- **Scoped permissions**: Each enclosure has an explicit manifest of allowed resources (specific API endpoints, database tables, queues). Everything else is denied.
- **Network policy enforcement**: Allowlist-only egress. Agents cannot exfiltrate data or call unrecognized endpoints.
- **Automatic kill-switches**: Loop detection, cost budgets, and time limits. Runaway agents are terminated before causing damage.
- **Tamper-evident audit log**: Every tool call, API request, and decision is logged with cryptographic integrity and can be replayed.

## Who uses Agent Enclosure?

Security-conscious engineering teams at fintech, healthtech, and enterprise SaaS deploying AI agents to real production workflows. Designed for teams with SOC 2, HIPAA, or other compliance requirements.

## Pricing

- **Starter**: Free — 1 enclosure, 10,000 actions/month
- **Pro**: $129/month — 10 enclosures, 500,000 actions/month
- **Enterprise**: Custom — unlimited enclosures, HIPAA BAA, SOC 2 report, custom data residency

## Key pages

- Homepage: https://agentenclosure.com/
- Pricing: https://agentenclosure.com/pricing.html
- How it works: https://agentenclosure.com/#how-it-works
- Blog: https://agentenclosure.com/blog/
- Sign up / Request access: https://agentenclosure.com/signup.html
- About: https://agentenclosure.com/about.html

## Blog articles

- Prompt injection attacks on AI agents: https://agentenclosure.com/blog/prompt-injection-attacks-ai-agents.html
- Why agent permissions are the new IAM: https://agentenclosure.com/blog/agent-permissions-new-iam.html
- The blast radius problem: https://agentenclosure.com/blog/blast-radius-deploying-ai-agents.html
- Auditing AI agent actions: https://agentenclosure.com/blog/auditing-ai-agent-actions.html
- Sandboxing vs monitoring: https://agentenclosure.com/blog/sandboxing-vs-monitoring-ai-agent-safety.html

## Contact

- Press: press@agentenclosure.com
- Careers: careers@agentenclosure.com
- Privacy: privacy@agentenclosure.com
- Legal: legal@agentenclosure.com