Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)

Yandex 360 Launches MCP Server, Giving AI Agents Access to 7 Office Services

Yandex 360 has launched an MCP server, a single "socket" through which ChatGPT, Claude or DeepSeek can work with the Yandex office suite without custom-built integrations. Seven services are connected so far: Mail, Disk, Calendar, Telemost, Wiki, Tracker and Forms. For example, an assistant can find a correspondence thread with a contractor and check it against tasks in Tracker. The agent acts with the employee's permissions and will not send an email on their behalf. It only prepares a draft.

AI-processed from TG: @ai_machinelearning_big_data (Machinelearning); edited by Hamidun News
Yandex 360 Launches MCP Server, Giving AI Agents Access to 7 Office Services
Source: TG: @ai_machinelearning_big_data (Machinelearning). Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTG: @ai_machinelearning_big_data (Machinelearning) ↗
Editor-in-Chief
Access is in preview for nowfor Yandex 360 customers and paid plans
Key metric #2With admin rights, the agent can add employees and change work groups
Key metric #3The server runs on Yandex's side, but customers set up the connection and the agents themselves

Until now, the AI assistant a team uses and the company's working data lived separately. For ChatGPT or Claude to see email, calendar and tasks, developers had to write integrations by hand. Yandex 360 removes that chore: the company has launched its own MCP server, and external AI agents can now access 7 services of the virtual office through a standard protocol.

MCP (Model Context Protocol) is an open protocol from Anthropic, a kind of universal socket: a service exposes a set of tools once, and any compatible AI application can use them. For example, you can ask an assistant to find a correspondence thread with a contractor and check it against tasks in Tracker. This is worth reading for anyone building internal AI assistants or preparing their own service to work with agents.

What exactly Yandex has opened up

Previously, working with Yandex 360 through AI was possible only via Yandex's own solutions: Alice AI for business or agents built on the Yandex AI Studio platform. Now there is a second route. The services can be linked to a company's own systems and used through external AI interfaces: ChatGPT, Claude and DeepSeek.

Seven services already support connection via MCP: Mail, Disk, Calendar, Telemost, Wiki, Tracker and Forms. Yandex promises that other services in the suite will gain support for the protocol over time.

What the agent will be able to do

The basic scenario is searching and analyzing data. Say you need to prepare a project summary: the agent finds emails with the contractor, checks them against tasks in Tracker and a document on Disk, and compiles a short report. Previously, a person assembled this chain by hand, switching between tabs.

There is also an administrative side. Through MCP, you can manage the organization itself in Yandex 360 for Business: add new employees or change the membership of work groups. The user must have the appropriate administrator rights to do this. For an IT department head, this is a chance to automate routine onboarding: on request, the agent creates the new person's account and adds them to the right groups.

Limitations and security

The main question when connecting an agent to corporate email is what damage it could do. Yandex's answer: agents do not have unrestricted access to company data. They work with the same permissions as the user and can use only the tools they are allowed. A telling example: the agent cannot send an email on a person's behalf. It can only write a draft, which the person reviews and sends themselves.

This is reasonable protection against the most unpleasant mistake, where a model sends clients something it shouldn't by accident. But it does not eliminate the problem entirely: everything available to an employee is available to their agent. If a person has broad permissions, the agent will see a lot. The practical takeaway: before launch, check what access the people who will connect agents have, and start with accounts that have narrow permissions.

There are also organizational caveats:

  • The server runs on the Yandex 360 side, but the connection and the agents themselves must be set up separately. There is no ready-made "switch on and it works" mode.
  • The server is currently in preview, so changes are possible.
  • Organizations that are Yandex 360 customers can connect, as can users on paid plans.

What this means for developers and founders

For developers, the main news is that an office suite now has a standard entry point for agents. If you already build internal assistants on Claude or ChatGPT, you can now connect email, calendar and tracker to them without custom wrappers, and choose the model to fit the task.

For founders and product managers, the signal is different: MCP is becoming the usual way to open a service to agents, and now a major office product is doing it. If you have a B2B service, think ahead about which operations you are ready to hand to agents and with what permissions. Yandex's approach of drafts instead of sending is a good model: reading and preparing actions are allowed, while irreversible steps stay with the human.

It is too early to say which scenarios will prove most popular: the service has only just been released in preview.

Story

Key Questions Answered

What exactly Yandex has opened up?

Previously, working with Yandex 360 through AI was possible only via Yandex's own solutions: Alice AI for business or agents built on the Yandex AI Studio platform. Now there is a second route. The services can be linked to a company's own systems and used through external AI interfaces: ChatGPT, Claude and DeepSeek. Seven services already support connection via MCP: Mail, Disk, Calendar, Telemost, Wiki, Tracker and Forms.

What the agent will be able to do?

The basic scenario is searching and analyzing data. Say you need to prepare a project summary: the agent finds emails with the contractor, checks them against tasks in Tracker and a document on Disk, and compiles a short report. Previously, a person assembled this chain by hand, switching between tabs. There is also an administrative side. Through MCP, you can manage the organization itself in Yandex 360 for Business: add new employees or change the membership of work groups.

What are the key details on Limitations and security?

The main question when connecting an agent to corporate email is what damage it could do. Yandex's answer: agents do not have unrestricted access to company data. They work with the same permissions as the user and can use only the tools they are allowed. A telling example: the agent cannot send an email on a person's behalf. It can only write a draft, which the person reviews and sends themselves.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Hands-On AI Engineering · Hamidun Academy

Build & Ship Production AI Agents Without Boilerplate Code

Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.

Start 72h Free Trial →Vibe-Coding Intensive
What do you think?
Loading comments…