Platform Features Modyo and AI Channels Customers Origination Content Insights Connect Dynamic

MODYO MCP SERVER

Build with your agents. Launch with Modyo.

Connect the AI agents your team already uses to Modyo, with your models and your rules. They build sites, widgets and flows with your platform's real context and the same permissions as the person using them: your team controls what AI accelerates.

Read the docs See how it works

  1. Your AI agent

  2. Modyo MCP Server

    • WIDGETS
    • CONTENT
    • PAGES
    • FORMS

  3. Your site in development

  4. Review

With the permissions of the person using it

FROM PILOT TO PRODUCTION

The question isn't what AI can do anymore. It's what can reach production.

The 2024 and 2025 pilots were for learning. In 2026, our clients expect real, high-value use cases in production, with control and traceability. The Modyo MCP Server is available today, with your AI on Modyo's foundation.

  • From pilots to results.

    AI-powered work in production with metrics you can defend.

  • Return where it matters.

    Time to market, IT workload and development cost.

  • AI-ready data.

    AI is only as reliable as the data behind it: structured, well-managed data is what moves a use case into production.

WHY MODYO

Why Modyo matters more in an AI world

Five reasons you can share with your leadership team.

  1. From pilot to production

    Your agents build with Modyo in a proven AI ecosystem: what they create reaches review ready to launch.

    See how it works

  2. Your AI, or ours

    The MCP Server connects to your team's AI ecosystem: the agents and models it already uses, such as Claude, OpenAI, Gemini, GitHub Copilot or Cursor. Modyo doesn't impose a model.

    See how it connects

  3. Your team controls

    The agent works with the permissions of the person using it and goes through the same approvals: nothing launches without your team.

    See the controls

  4. A proven Modyo foundation

    The MCP Server works on Modyo's APIs, which serve mission-critical channels and provide a stable foundation.

  5. Content agents understand

    When an agent asks, your site stops being the interface and becomes the source. Modyo's structured content helps assistants describe you accurately.

TWO WAYS TO BUILD

Two ways to build with your agents

Your team keeps its tools; the agent brings Modyo's context.

  • Sites and content.

    Pages, widgets, content and forms, built with your platform's real context.

    • Ask in plain language
    • With your platform's real context
    • With your team's conventions
    Watch it get built

  • Financial micro frontends.

    Banking and insurance experiences on the Modyo Dynamic framework and design system, with harness engineering operated by Modyo.

    See Modyo Dynamic

Prefer that we run it? Modyo AI Workflows runs your AI workflows as a managed service, with your models and your rules.

Request a demo

Read how to set up the MCP Server in the docs

HOW IT WORKS

Four steps, with your team in charge

Your agent does the heavy lifting. Your team decides what gets published.

Read the docs

  1. Connect your agent.

    Add the Modyo MCP Server to your AI agent with an API token that carries the permissions of the person who creates it.

  2. Describe what you need.

    Ask in plain language for a widget, a page, a form or an origination flow.

  3. The agent builds.

    It works in your development space with Modyo's tools, resources and conventions.

  4. Your team reviews and publishes.

    You review it, approve it and launch it with the same roles, approvals and audit as always.

Control

Control stays with your team

AI without control is a risk. In Modyo, the agent works with the permissions you give it, logs every action, and nothing launches without your team.

  • One person's permissions.

    The agent acts with the access of the person using it, across widgets, pages, templates and content, and never with more.

  • The approvals you already have.

    If that person needs approval to publish, so does the agent. Your approval workflows don't change.

  • Full traceability.

    Every agent action is recorded in Modyo's standard audit logs, under the account of the person using it.

The MCP Server works while you build and adjust your experiences, not in the live operation of your channels.

  • Team review before launch
  • Roles and permissions
  • Approval workflows
  • Versions
  • Audit logs
  • ISO 27001
  • CSA STAR Level 2
  • SOC 2 Type II report

Proof

Experience you can verify

years building critical digital channels

Security seals: ISO 27001, AICPA SOC and CSA STAR Level Two
  • ISO 27001
  • CSA STAR Level 2
  • SOC 2 Type II report

Building their digital channels with Modyo

  • Banco BICE
  • Santander
  • BCI
  • Banco de Chile
  • Consorcio
  • Mibanco
  • SURA
  • Invex
  • Mifel
  • LarrainVial
  • SBS
  • Copec
  • Caja Los Andes
  • Parque Arauco
  • CFG
  • Duoc

FAQ

Frequently asked questions

MCP (Model Context Protocol) is an open standard that lets AI agents use software tools. The Modyo MCP Server exposes Modyo's capabilities as tools, resources and prompts, so your agents build inside your platform with your permissions and your rules.

Widgets, pages, content and content types, forms and origination flows. It can also check publishing status, versions and approval workflows.

The agent can only do what the person using it can do, because it works with that person's roles and permissions, across widgets, pages, templates and content. If your team requires approval before publishing, that rule applies to the agent too.

Any MCP-compatible agent, such as Claude, OpenAI, Gemini, GitHub Copilot or Cursor.

Yours. The MCP Server connects to the agents and models your team already uses, so the AI works under your rules, not ours.

Yes, with Modyo Dynamic: Modyo's framework and design system for financial micro frontends, with harness engineering operated by Modyo.

A Modyo managed service to run AI workflows under your rules. If you use your own model, it connects to your account on AWS Bedrock, Google Vertex, Azure AI Foundry or another compatible provider, and inference runs under your contracts. Request a demo to see it.

An Enterprise plan on Modyo Platform 10 or later, an API token with the right permissions and an MCP-compatible agent. Our team runs enablement sessions for the initial setup.

The Modyo MCP Server is available to customers on the Enterprise plan. Request a demo and we'll show you how to get started.

No. The MCP Server works while you build and adjust your experiences; live operation stays with your team and the platform.

Resources

Take your AI from pilot to production

Request a demo and we'll walk through how the MCP Server fits your channels, your models and your rules.

Request a demo Read the docs