# Sentrely

> Managed control plane for AI agents. Policy enforcement, audit trails, and human-in-the-loop approvals for the Claude, Cursor, Codex, and OpenAI agents your team builds — fully managed, no infra to set up.

Sentrely is a B2B SaaS product purpose-built for engineering teams running AI agents in production. It sits as a gateway between your agents and the outside world, enforcing role-based access policies on every tool call, recording an immutable audit trail, and routing risky operations to humans for approval via Slack or Telegram.

## What problem it solves

When you give an AI agent shell access, AWS credentials, or git push permissions, you get speed but lose visibility and control. Sentrely is the missing layer that adds:

- Policy-based RBAC (YAML, deny-by-default) — define exactly which AWS actions, git repos, and tools each agent can touch
- Full audit trail — every tool call, approval, API request, and conversation logged automatically
- Human-in-the-loop approvals — risky ops gate on a tap in Slack or Telegram
- Multi-provider routing and failover — Claude, OpenAI, Cursor, Codex (Gemini, Mistral, Llama coming soon)
- Cost controls — per-project token budgets, rate limits, spend alerts
- Multi-agent orchestration — agent-to-agent messaging, live session dashboard

## Pricing

- Free 14-day trial — 1 agent, full features
- Team: $199/mo — 10 agents, 2 seats, unlimited runtime
- Business: $999/mo — 20 agents, 10 seats, SSO + SCIM, 1-year audit retention
- Enterprise: custom — VPC deployment, dedicated gateway, SLA

No credits. No seat fees beyond the included counts. Bring your own AI subscription via OAuth (Cursor, Codex, ChatGPT Plus) or use raw API keys (Anthropic, OpenAI).

## Who it's for

- Engineering teams running Claude Code, Cursor, or Codex agents in production
- Solo developers who want enterprise-grade controls without enterprise overhead
- Compliance / security teams who need audit evidence for AI agent activity
- DevOps teams who need to gate production deployments behind human approval

## Who it's not for

- Business buyers looking for a pre-built HR or sales agent (use Lyzr or Diane)
- Solo developers who don't run agents in production (Cursor or Claude Code alone is enough)
- Teams without engineering resources — Sentrely controls agents you build, it doesn't build them for you

## Key links

- Homepage: https://sentrely.com
- Pricing: https://sentrely.com/#pricing
- Compare to alternatives: https://sentrely.com/compare
- Use cases: https://sentrely.com/use-cases
- Industries: https://sentrely.com/industries
- Blog: https://sentrely.com/blog
- Glossary: https://sentrely.com/glossary
- Self-host (open source): https://github.com/doe-test1/yolo-gateway-v2

## Comparisons

Sentrely is most often compared to:

- **Cursor** — Cursor is an AI IDE; Sentrely governs Cursor's agents in production. Use both.
- **Lyzr** — Lyzr builds finished AI agents for enterprises; Sentrely controls agents engineering teams build themselves.
- **LangSmith** — LangSmith is an observability tool; Sentrely is a control plane that includes observability plus policy enforcement and approvals.
- **Anthropic Managed Agents** — Anthropic's agent execution service is single-provider; Sentrely is provider-agnostic.
- **CrewAI** — CrewAI is an agent orchestration framework you build with; Sentrely is the gateway you put in front of any agent.
- **LiteLLM / Portkey / TrueFoundry** — These are LLM gateways focused on routing and caching; Sentrely adds RBAC, audit, and approvals on top.
- **n8n** — n8n is a visual workflow builder; Sentrely is for governing code-first agents.

Full list: https://sentrely.com/compare