SI Agent News
DevelopingAgentsSecurity

31C lets AI agents query its network-traffic analysis tool through MCP

31C has opened its ODUN.ONE network-analysis product to AI agents using the MCP standard, in private beta. It is a vendor claim, but it shows where agent plumbing is heading.

2024year MCP, the open standard behind this launch, was introduced

The short version

Why it matters

Network data has usually been locked behind specialist teams and custom integration work. If standards like MCP make that access routine, more AI agents will touch sensitive operational data, and the real question becomes what each agent is allowed to see and do. For now this is a company announcement, so treat its benefits as claims.

What 31C announced

31C, a company based in Dubai, has launched MCP ONE. As Yahoo Finance reported in a press release, it gives AI software agents a direct way into ODUN.ONE, the company's platform for analyzing network traffic. ODUN.ONE already had a screen for people and a programming interface for other machines. Agents are now the third kind of user.

The product is available now in private beta, meaning a limited group of customers is testing it. The release does not give a price, a customer list or performance figures.

What MCP is and how it fits

MCP stands for Model Context Protocol. It is an open standard, introduced in 2024, that gives AI agents a common way to reach tools and data. The release says the major AI platforms now support it. Think of it as a universal plug: build the socket once, and any agent that has the matching plug can connect.

ODUN.ONE is described as a deep packet intelligence platform. That means it reads the traffic moving between an organization's users, applications and services and works out what is happening. The company says an agent can get a structured answer in real time, even when the traffic is encrypted.

The setup works like this. A customer deploys MCP ONE once on its own infrastructure. After that, 31C says, connecting a new agent just means telling it to use MCP ONE, with no custom connector. The agent can act on what it learns only within the permissions the customer sets.

The problem 31C says it removes

According to the release, the record of what moves across a network is among the richest data a large company or network operator holds. Until now it sat behind a specialist team. A department that wanted it had to file a request, wait weeks for someone to build and tune an integration against the existing API, and wait again whenever something changed. In 31C's telling, most agents never got the data at all.

MCP ONE is pitched as ending that wait. The company describes the integration as frictionless and says one instruction replaces a weeks-long project. That is the vendor's own framing, and the release offers no independent test.

Four examples from the release

The company sketches what a normal Monday could look like:

In the security example, the agent proposes and the analyst decides. That detail matters, because it shows where the human stays in the loop.

How to read this announcement

This is a press release, not a review or a customer report. It shows that a specialized product has adopted a general standard. It does not show that the product works as described, or how well. The company's claims about speed, ease and unique data have not been checked by anyone outside it.

The more useful signal is the pattern. MCP began as a way to connect chat assistants to tools such as files and calendars. Here it appears in a corner of enterprise networking, where the data is sensitive and the stakes of a wrong action are higher.

What to watch next

The private beta is the first test. Useful evidence would be named customers, published pricing and clear detail on how permissions are set and logged. The release says agents act within customer-set permissions, but it does not say how fine-grained those controls are.

Readers who run networks or security tools can ask their own vendors whether they offer an agent connection. Before saying yes to any, decide what an agent should be allowed to read and what it should only suggest.

What to watch

Try it

If your organization uses network or security tools, ask each vendor whether it supports MCP and what permission controls an agent connection would have.

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by Yahoo Finance · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news