Runlayer puts one gateway between agents and MCP servers and ties every call to a person through SSO and SCIM. Policies apply per user, agent and tool call, risky actions can require human approval, and it finds unmanaged MCP servers and client configs on managed devices through your existing MDM. It is a full enterprise platform rather than a proxy you drop into a repo.
Who it suits
Security and platform teams rolling MCP out across a whole company, with SSO, tool-level policy and shadow MCP discovery in one place.
Strengths
- Identity on every call: SSO, SCIM and just-in-time credentials, so each tool call is attributed to a user and an agent.
- Policy at tool-call depth, by user, agent, client and tool, with human approval for risky actions.
- Shadow MCP discovery on managed devices through your MDM, without employee opt-in.
- Runtime scanning for tool poisoning, prompt injection, intent drift, PII and exfiltration, starting in alert-only mode before blocking.
- Built-in OAuth and credential handling, a catalog of approved MCP servers, audit trail and OpenTelemetry tracing.
Limits
- An enterprise platform, not a lightweight open-source proxy: there is no repo to clone and run on a laptop.
- Pricing is not published; you get it through a demo.
- Self-hosting options are not described on the public site; ask if you need to run it in your own environment.
Scores on each criterion
| Criterion | Score | What we found |
|---|---|---|
| Authentication | Strong | SSO and SCIM, with user and agent attribution on each call. |
| Credential handling | Strong | Built-in OAuth and credential handling; just-in-time credentials. |
| Policy depth | Strong | Policies per user, agent and tool call; human approval for risky actions. |
| Audit | Strong | Full audit trail with user and agent attribution; tamper-proof logs. |
| Observability | Strong | Session traces and OpenTelemetry tracing; usage and spend views. |
| Deployment model | Partial | Managed platform; self-hosting not described publicly, so ask. |
| Shadow usage discovery | Strong | Finds unmanaged MCPs, skills and client configs via MDM. |
| Total | 13/14 | Ranked #1 of 8 |
Strong Built in and documented (2 points)Partial Present with gaps, or limited scope (1 point)Not documented Not found in the public docs we reviewed (0 points)
What it costs
Not published. Request it through runlayer.com/book-a-demo.
Sources
Every score above comes from these public pages. Check them for changes since 11 November 2026.