# Mintive

> Enterprise platform that governs AI agents' access to tools, models, and data through MCP and LLM gateways, agent identities, monitoring, and coworker agents.

Mintive, formerly MintMCP, builds infrastructure for the agentic enterprise. It connects AI agents to models, tools, and company data through one platform that governs access by role and records every action. The platform covers an MCP gateway, an LLM gateway, agent identities, an agent monitor, and coworker agents in Slack and Teams.

## What It Is

Mintive is an AI agent governance platform. Its premise is to treat each agent like an employee: it gets its own identity, only the access its work requires, and a record of what it did, rather than borrowing a person's credentials and inheriting all their permissions. One set of rules is meant to cover Claude, ChatGPT, Copilot, Cursor, and agents that teams build themselves.

## Gateways and Access Control

The MCP gateway routes AI tools through a single gateway. Teams approve which MCP servers can be used, group them into role-based bundles, and switch off destructive tools at the tool level. A custom MCP registry lets engineering curate a catalog of approved servers. The LLM gateway routes model calls from tools such as Claude Code and Codex, setting allowed models and budgets per team and supporting providers such as OpenAI, Anthropic, Fireworks, and OpenRouter.

## Identities, Monitoring, and Protection

Agent identities give each agent delegated credentials and least-privilege access that can be revoked individually. Every call is traced to a person, agent, server, and model, with spend and token breakdowns by team; audit logs can be exported to a SIEM or streamed via OTLP. Mint Guard screens tool calls for prompt injection, exposed secrets, personal data, and harmful content, and private tunnels reach servers inside a company network.

## Coworker Agents and Compliance

Coworker agents work in Slack and Teams with long-term memory and the ability to take on work proactively, using the same governed access. Mintive states it is SOC 2 Type II audited, supports enterprise SSO via SAML/OIDC, offers EU and US data residency, and signs BAAs for customers handling protected health information.

## Features
- MCP gateway with approved servers and role-based bundles
- Tool-level switches to disable destructive actions
- Custom MCP registry
- LLM gateway with per-team budgets and allowed models
- Agent identities with scoped, revocable credentials
- Agent monitor with call tracing and cost breakdowns
- Audit log export to SIEM or via OTLP
- Mint Guard screening for prompt injection, secrets, PII, and harmful content
- Private tunnel to servers inside company networks
- Coworker agents in Slack and Teams with long-term memory
- SSO via SAML/OIDC
- EU and US data residency

## Integrations
Claude, ChatGPT, Cursor, Copilot, Codex, Slack, Microsoft Teams, Okta, Gmail, Outlook, Google Calendar, Linear, Notion, Snowflake, Elasticsearch, Salesforce, OpenAI, Anthropic, Fireworks, OpenRouter

## Platforms
WEB, API

## Pricing
Paid

## Links
- Website: https://www.mintive.com
- Documentation: https://www.mintive.com/docs/intro
- EveryDev.ai: https://www.everydev.ai/tools/mintive
