Control and build Frends integrations straight from your AI client

Connect once. Build integrations, update processes, debug issues and reason over your entire platform, all through plain language in Claude, ChatGPT or any MCP-compatible AI. With every change governed and human-approved before it goes live.

Enterprise MCP Hero

Enterprise teams are moving fast on AI. Integration workflows can now keep up.

Your team works in AI tools. But every time they need to build, update or troubleshoot an integration, they have to stop and switch to the Frends UI. Delivery has never been this fast.

Building and updating integrations means leaving your AI environment

Developers spend 40 to 60% of their time navigating through the UI, searching for the right tasks and manually translating what their AI assistant suggested into platform actions. Now they get this time back.

Debugging a broken integration still means finding the original developer

When something goes wrong in production, you need someone who knows the process. That means escalations, delays and blocked queues, even for simple diagnostic questions. Get ready to solve issue within minutes.

AI-to-platform control with no need for custom integration work

Most teams that want AI to interact with their integration platform end up writing bespoke scripts, managing authentication manually and maintaining a fragile layer between their AI tools and their iPaaS. There's a better way.

Frends MCP covers the full DevOps loop. All from your AI client

Frends MCP Server makes every Frends platform capability accessible through conversation. Forget raw API wrappers. This is a purpose-built capability for AI consumption, with a set of intelligently designed tools that let your AI client perform complex work reliably in far fewer steps than a generic integration approach.

 

Connect your AI client once. Then build, update, monitor, debug, and reason over your platform without ever leaving the environment you work in.

Key capabilities

What you can do with Frends MCP

Build and update integrations through conversation

Describe a new integration or a change to an existing one in plain language. Frends searches the task catalog, configures the right connectors and returns a complete BPMN 2.0 process draft, ready to review and deploy in the Frends UI.

Monitor, debug and reason over your platform

Ask what's broken, what's slow, what ran last night or what would be affected if you changed a system. Frends MCP gives your AI client direct access to execution history, process metadata and platform-wide dependency maps.

Search and discover your process catalog

Find existing processes by name, tag, system or behaviour through natural language. Understand what you already have before building something new. Reuse proven patterns instead of rebuilding from scratch.

Full governance & access control from day one

Every MCP action is tied to a specific user via personal non-shareable tokens. Access is scoped to each user's permissions. Nothing goes live without review and approval in the Frends UI. Every tool invocation is logged.

See Frends MCP in action

Take the interactive demo to see how Frends MCP works, from describing an integration in plain language to watching it build a process draft; from asking what's failing in production to getting a step-by-step diagnostic.

What this means for your team

Here's a preview of what Frends MCP enables:

Modernize legacy systems with AI access in hours-2

Process creation and updates that previously required UI navigation, task searching and manual configuration now happen through a single prompt. From requirement to deployable draft in minutes.

Deploy AI with confidence in regulated industries.-2

A purpose-built MCP implementation means your AI client completes complex tasks reliably and efficiently. In one real scenario, AI querying raw data tables directly consumed approximately 10 million tokens for a single query. The same result through Frends MCP: a few thousand tokens.

Connect your AI of choice to any custom or legacy system-2

"Our order confirmations aren't reaching customers, can you find out what's going on?" Your AI client traces the integration, inspects execution history and identifies the failure. What previously required a specialist now takes minutes through conversation.

Keep what you’ve built, and make it AI-ready-2

"We're upgrading our ERP to the new API version, can you check what's affected, and how big is this?" Frends MCP maps your entire dependency landscape through conversation. No manual inventory. No guesswork.

Run your AI tools anywhere_ cloud, on-premises or boht.-2

New to a process? Inherited someone else's integration estate? Ask Frends MCP what a process does, how it's structured, what it connects to, when it last ran. The knowledge that used to live in one person's head is now one question away.

Modernize legacy systems with AI access in hours-2 Deploy AI with confidence in regulated industries.-2 Connect your AI of choice to any custom or legacy system-2 Keep what you’ve built, and make it AI-ready-2 Run your AI tools anywhere_ cloud, on-premises or boht.-2

Make your systems AI-callable with Enterprise MCP

Frends MCP is not only about prompting integrations and letting your team control Frends through AI. With Enterprise MCP, any system can become AI-callable through Frends processes. So your AI agents can reach legacy systems, custom ERPs and internal applications without touching a line of existing code.

Use cases

See it working in the real world

Integration developer
Integration developer
Platform operations
IT planning
Financial processes

Four processes built before the coffee got cold

A solution architect received a customer's PoC specification on a Friday. Using Frends MCP from Claude, he described each integration in plain language, reviewed the generated BPMN process drafts and had all four running in the environment before the team returned on Monday.

Custom & legacy systems

Process run failure diagnosed in minutes

A production process failed overnight. An operations engineer asked Frends MCP: "Our salary processes didn't run last night — what happened?" The AI traced the execution history, identified a failed dependency, and returned a clear diagnostic with a link to the affected process in the UI. The issue was resolved before anyone escalated.

Finance (invoice handling)

System replacement impact mapped through conversation

An IT manager preparing to replace a CRM system asked: "What Frends processes connect to HubSpot, and what would be affected if we moved to Pipedrive?" Frends MCP returned a full dependency map, every connected process, its execution frequency and which business functions it served. A project that previously required a manual audit took minutes.

Finance (credit risk)

Bottlenecks surfaced without a business analyst

A finance operations lead asked: "Do we have any bottlenecks in our month-end financial processes?" Frends MCP reviewed execution history and process timing across the relevant workflows, identified that the reconciliation process was taking several days and proposed a targeted integration to automate the slowest step.

Finance (credit risk)-1

Webinar: Introducing Frends MCP

Watch the Frends MCP Server live demo and see how your team can build integrations, debug production issues and reason over your platform — all through conversation — in a session presented by Frends CTO Asmo Urpilainen.

FAQ

What is Frends MCP Server?

The Frends MCP Server is a capability built into the Frends platform that makes every iPaaS function — process creation, search, execution, monitoring, and debugging — accessible through any MCP-compatible AI client, including Claude, ChatGPT, Copilot and Cursor. It uses the Model Context Protocol open standard and exposes 57 dedicated tools purpose-built for AI consumption. Unlike a raw API wrapper, each tool is designed for AI reasoning, so your AI client can complete complex, multi-step tasks reliably and efficiently.

How is Frends MCP different from just connecting an AI to the Frends API?

The difference is significant. A raw API connection requires the AI to compensate for documentation gaps, pagination logic, authentication handling and error patterns on its own, which consumes enormous amounts of tokens and frequently fails on complex tasks. Frends MCP provides tools that are specifically designed for how AI reasons: structured, outcome-oriented and optimized to minimize the number of steps needed to complete a task. In one real scenario, a raw AI-to-data approach consumed approximately 10 million tokens for a single query. The same result through Frends MCP: a few thousand tokens.

What can I build or do through Frends MCP?

You can create new integration processes from plain language descriptions, update existing processes, search and inspect your process catalog, trigger manual executions with parameters, check execution status and history and ask diagnostic questions about your platform, all from inside your AI client. The full list of capabilities expands with each release.

Does this work with Claude, ChatGPT and other AI clients?

Yes. Any MCP-compatible AI client can connect to the Frends MCP Server. This includes Claude Desktop, ChatGPT, Cursor and custom AI agents. The connection requires a personal access token generated from your Frends account.

Does anything go live automatically?

No. Every process created or updated through Frends MCP is created as a draft in the Frends UI. A human reviews and approves it before it is deployed. Nothing goes live automatically as this is a core design principle.

How does authentication and access control work?

Each user generates a personal, non-shareable access token in their Frends account. This token scopes all MCP access to that user's permissions, they only see the tools and processes they are authorized to access. Administrators cannot view user tokens. Tokens are automatically invalidated when a user is deactivated. Every tool invocation is logged with user identity, tool called, parameters, result and timestamp.

Can Frends MCP run on-premises or in air-gapped environments?

Yes. The Frends MCP Server runs within the Frends UI service, which can be deployed fully on-premises, in a private cloud or in an air-gapped environment. This makes it suitable for regulated industries, such as finance, energy, defense and healthcare where AI processing cannot leave the organization's own infrastructure.

How does Frends MCP relate to Enterprise MCP?

They are two sides of the same capability. The Frends MCP Server lets your team control the Frends platform through AI, by building, updating, monitoring and debugging integrations via conversation. Enterprise MCP covers the other side of the equation: any system can become AI-callable through Frends processes, giving AI agents governed access to legacy systems, custom ERPs and internal applications without the need to code anything. Most organizations will use both together.

Is there an additional cost for Frends MCP Server?

The Frends MCP Server is included in Business and Enterprise tiers. Each MCP tool invocation counts as one process execution against your existing monthly limit. There is no additional licensing fee.

Ready to control Frends from inside your AI client?

See how Frends MCP covers the full DevOps loop — from building and updating integrations through conversation, to monitoring and debugging your platform in production — in a live, personalized demo with our team.