Preloop vs MintMCP: Open-source MCP gateway with built-in approvals

Preloop is the open-source AI agent control plane — an MCP firewall, AI model gateway, human approvals, runtime observability, and audit trail in one self-hostable platform. MintMCP is commonly positioned as a SaaS MCP gateway with enterprise human-in-the-loop approvals integrated into Slack and Teams. The honest one-liner: if you want a plug-and-play SaaS MCP gateway, MintMCP is a credible choice; if you want an open-source, self-hostable control plane that covers MCP governance plus model traffic, approvals, and audit evidence, choose Preloop.

At a glance

What MintMCP does well

MintMCP is generally described as a thoughtful SaaS MCP gateway that emphasizes enterprise-ready human-in-the-loop for MCP tool calls:

• Fast time-to-value for SaaS buyers. If you want to put an MCP gateway in front of Claude Desktop, Cursor, or other MCP clients quickly, a hosted gateway minimizes setup.
• HITL approvals in Slack and Teams. Reviewers can approve or reject tool calls in the collaboration surface they already use.
• Enterprise positioning. The product is targeted at organizations that want MCP governance without running infrastructure.
• Focused scope. MintMCP keeps its surface area centered on the MCP layer, which can simplify evaluation for teams only shopping for an MCP gateway.

If your only requirement is a SaaS MCP gateway with Slack/Teams approvals, MintMCP is a reasonable option.

Where Preloop is different

Preloop overlaps with MintMCP on MCP governance and HITL approvals, but the scope and shape of the product are different.

• Open source and self-hostable. Preloop's core is Apache 2.0 licensed. You can run it on your own infrastructure, read and audit the code, and adapt it to your environment. There is no requirement to send MCP traffic to a third-party SaaS.
• Bundled model gateway. Preloop is not only an MCP firewall. It also includes an AI model gateway that is OpenAI/Anthropic-compatible, with per-agent, per-flow, and per-key budgets and attribution. One product governs tool calls and model traffic.
• Runtime session observability. Drill from a fleet view into any single agent session and see the full timeline of tool calls, model calls, matched policies, approvers, and outcomes. This is operational evidence you can use for incident review or AI governance programs.
• Policy-as-code with CEL. Allow / deny / require-approval / require-justification rules live in YAML with CEL expressions for fine-grained conditions on tool arguments. Ordered rules and priority give you predictable behavior, and policies version-control like the rest of your infrastructure.
• preloop agents discover for local coding agents. One command enrolls Claude Code, Codex CLI, Cursor, Gemini CLI, OpenClaw, and OpenCode on a developer's machine by rewriting their local configurations — no SDK changes, no per-repo work.
• Approval channels beyond Slack/Teams. Mobile and watch push, Slack, Mattermost, email, and custom webhooks. Async-safe mode lets long-running reviews complete without breaking the agent's transport.

When to choose MintMCP

• You want a hosted, SaaS-only MCP gateway without operating any infrastructure.
• Your approval surface is strictly Slack or Teams and you do not need mobile-first or webhook approvals.
• You do not need a model gateway or runtime session observability in the same product.

When to choose Preloop

• Platform and DevEx teams rolling out AI coding agents (Claude Code, Codex CLI, Cursor, Gemini CLI, OpenClaw, OpenCode) who need zero-touch onboarding via preloop agents discover.
• Security teams in regulated SaaS that require a self-hostable control plane and searchable audit evidence that lives inside their own network.
• Engineering teams standardizing on one control plane for MCP tool governance, model traffic, and approvals — instead of buying a gateway here, an approvals SaaS there, and an LLM observability product separately.
• Open-source-first organizations that want a control plane they can read, fork, and extend.
• Teams building EU AI Act readiness evidence who need durable logs of matched policies, approvers, and outcomes.
• Organizations with cost-governance requirements that want model budgets and attribution natively inside the MCP control plane.

Adoption notes

If you are already using MintMCP for MCP approvals, Preloop can be introduced side-by-side: start with preloop agents discover on a developer machine to pipe Claude Code, Codex CLI, or Cursor through Preloop for a test flow, and keep existing MCP traffic routed through MintMCP for other clients. Runtime sessions and audit evidence from the Preloop pilot can be compared against your existing approvals data. When you are ready, point the remaining MCP clients at Preloop's MCP endpoints and decommission the external gateway.

Because Preloop is open source, you can also run it inside a private VPC or on-prem subnet for sensitive environments where sending MCP traffic to a third-party SaaS is not acceptable.

FAQ

Is there an open-source alternative to MintMCP? Yes. Preloop is an open-source AI agent control plane released under Apache 2.0. It includes an MCP firewall with policy-as-code, human-in-the-loop approvals, an AI model gateway with budgets, runtime observability, and an audit trail.

Can I self-host my MCP gateway? Yes, with Preloop. The core is Apache 2.0 and runs on your own infrastructure. You keep MCP traffic, logs, and approvals inside your network.

Does Preloop support Slack and Teams approvals? Yes. Preloop sends approval requests to Slack, Mattermost, email, custom webhooks, and mobile/watch push notifications. Approvals carry full tool arguments and agent reasoning so reviewers have context.

How do MCP tool policies work in Preloop? Policies are YAML files with ordered rules. Each rule can match on tool name, server, parameter values, agent identity, or other context via CEL expressions, and decide allow, deny, require approval, or require justification.

Does Preloop cover model traffic in addition to MCP? Yes. Preloop ships with an OpenAI/Anthropic-compatible model gateway that tracks spend per agent, flow, and API key, with budgets and attribution. You get MCP governance and LLM traffic governance in one product.

How do I onboard existing agents to Preloop? Install the Preloop CLI and run preloop agents discover. Preloop finds local Claude Code, Codex CLI, Cursor, Gemini CLI, OpenClaw, and OpenCode configurations, backs them up, and rewrites them to route through Preloop.

Try Preloop

• Start your free trial
• Book a demo
• Preloop on GitHub