Mintive Raises $7m to Give AI Agents Identities

Mintive, the start-up formerly known as MintMCP, has raised $7m in a round led by Cowboy Ventures to build what it calls infrastructure for the agentic enterprise. The company announced the round in a launch post on its blog on 29 October, alongside its new name, and took its total funding to $11m.

New investors Operator Collective and Burst Capital joined the round, Mintive said, with existing investors Coatue, Hustle Fund and Maven Ventures taking part. Its angel investors include Okta co-founder Frederic Kerrest, WorkOS founder and chief executive Michael Grinich, former Adobe chief product officer Scott Belsky, former Google chief scientist Jeff Dean and OpenAI co-founder Andrej Karpathy. TechCrunch described the $7m as a seed round. Mintive did not disclose a valuation.

“Every AI agent at work runs with the access of whoever set it up,” chief executive Jiquan Ngiam wrote on X. “Today we’re launching @mintive: every agent gets its own identity and permissions, the same way every employee does.” The company’s own X account repeated the funding figure and its customer numbers.

What Mintive sells

Mintive’s idea is to treat each AI agent the way a company treats a new employee. In the launch post, Ngiam says each agent “gets its own identity, only the access its work requires, and a record of everything it does, regardless of the model or tool underneath”.

The company sells five products that work together:

MCP Gateway governs which tools agents can reach. MCP, the Model Context Protocol, is the common standard AI assistants use to connect to apps and data. Mintive puts every connector “behind one door”, with role-based bundles and switches for individual tools.
LLM Gateway routes calls through approved AI models, sets budgets by group and applies one policy across providers.
Agent Identity gives each agent its own credential, scoped to what it needs and revocable, with access delegated from the person it works for.
Agent Monitor records the chain behind every action: the person who started it, the agent that acted, the server it reached and the model involved. Teams can also track spending by user, model and tool.
Coworker agents work alongside staff in Slack and Teams with long-term memory. They are let into rooms explicitly and answer within the access of the person asking.

Mintive says the platform works with Claude, ChatGPT, Copilot, Cursor or a company’s own agents. A feature it calls Mint Guard screens calls for prompt injection, exposed secrets and personal data. The company says it is SOC 2 Type II audited and offers data residency in the EU and the US.

Customers and usage

Mintive says it has more than 50 paying customers, including Ro, Modern Treasury, Braze, Coursera, the University of Michigan, Apax and Stability AI. It says it processes more than 2.5 million agent calls a day and has handled more than 200 million in seven months. TechCrunch reported that the product launched in February. None of these usage figures has been checked independently.

The company gives two examples. At Modern Treasury, a bank operations agent searches tools such as Sentry, Notion, Linear and the codebase before a technical account manager steps in. Mintive says it went live in under three weeks and saves those account managers 30 minutes to four hours per case. At the University of Michigan, Mintive runs as the central MCP gateway for 150,000 students, faculty and staff.

Christian Burrows, head of security at Stability AI, is quoted in the post as saying that “simply blocking access isn’t an effective security strategy” because “people will find ways around it”.

Mintive also cites a Gartner forecast that the average Global Fortune 500 company will run more than 150,000 AI agents by 2028, up from fewer than 15 in 2025.

Why agent identity has become a crowded market

Mintive is small, but it is pushing into one of the busiest corners of enterprise AI. The problem it describes is real: an agent connected to company systems usually inherits all the access of whoever set it up, so one employee’s agent can see far more than its task needs. Ngiam argues that this is why security teams keep saying no.

Larger companies are chasing the same fix. IBM has made agent identity generally available and is previewing a kill switch for AI agents. Workday agreed to buy C1.ai, an identity platform for managing what people and AI agents can access. Salesforce has added agent evaluations and spending caps to Agent Fabric, and ServiceNow has formed an 11-member AI alliance with OpenAI and Anthropic, including Okta, to certify AI tools for governance.

Mintive’s pitch against those suppliers is neutrality. It argues that the controls on what an agent can reach should sit with the company, outside any single model or software vendor. That is a sensible argument for a firm using several AI tools at once. It also means Mintive has to persuade buyers to add one more layer at a time when their existing identity and IT suppliers are building the same features into products they already pay for.

A $7m round is modest next to the sums flowing into AI. Cowboy Ventures founder Aileen Lee, quoted in Mintive’s post, says the “major gating factors” for enterprise AI are governance, oversight, identity management and security, and that Mintive offers “trusted third party scaffolding” for them. The next test is whether a company of Mintive’s size can hold that position as bigger platforms close the gap.