Responsible AI in OpenProject

Where teams and AI agents work together.
Empower your team with AI while keeping decisions, data, and governance in your hands.

Get started with MCP

Artificial Intelligence in OpenProject

AI in your projects, on your terms

Bring AI directly into your project workflows with the OpenProject MCP Server. AI assistants and agents can work with current project context while your team stays in control of goals, decisions, and outcomes.

Built on the open Model Context Protocol (MCP), OpenProject connects compatible AI tools to projects, work packages, assignments, and other project information. Teams can ask questions, summarize progress, and let agents create or update project data within the permissions and tools you define.

Choose the AI client, model, and infrastructure that fit your requirements. OpenProject provides the connection and governance layer without locking you into a specific AI provider.

AI-powered project management that stays open, governed, and human-led.

AI project management

Connect OpenProject to the AI tools you choose

Use the OpenProject MCP Server as the secure bridge between your project data and MCP-compatible AI assistants and agents.

Your projects in OpenProject

Projects, work packages, assignments, schedules, and other project context remain in OpenProject, governed by your existing roles and permissions.

Your AI assistants and agents

Connect the MCP-compatible client and model that fit your organization.

  • Ask questions and explore project context
  • Summarize progress, risks, and open work
  • Create and update project information within permitted workflows
  • Choose where AI processing takes place

Get started with MCP

OUR PRINCIPLES

Responsible AI at OpenProject

Purpose-driven

AI that solves real project problems. We add AI where it can reduce routine work, surface relevant context, improve productivity, and support better decisions — not simply to follow trends or add complexity.

Human-led & Transparent

AI assistants and agents can support your team and act within defined workflows, while people remain accountable for goals, decisions, review, and outcomes. Users should understand when AI is involved and be able to validate its work.

Secure & Sovereign

OpenProject applies your existing security, roles, and permissions to MCP access. You decide which AI services can access project data and where that data is processed.

Open & Under Your Control

Use AI on your terms. The OpenProject MCP Server is based on an open standard, lets you choose your AI client and model, and gives your organization control over whether AI connectivity is enabled and which MCP capabilities are available.

Human expertise, amplified by AI

Teams and AI agents working together

AI agents can take on repetitive project work: finding information, summarizing status, preparing updates, and creating or updating project data. They work with the same project context your team uses in OpenProject.

People remain responsible for priorities, decisions, and outcomes. OpenProject provides guardrails for human-agent collaboration:

  • Agents work within defined permissions and available MCP tools.
  • Teams choose which AI client and model to use.
  • Organizations decide where data is processed.
  • People can review AI-supported work and remain accountable for project outcomes.

Illustration showing project work packages supported by AI.

Secure AI in OpenProject

Secure by design: OAuth2 with user context

Security is built into the connection. The OpenProject MCP Server uses OAuth2 authentication so interactions between an AI assistant or agent and OpenProject happen in the context of the authenticated user.

The connected AI can only access or change information through the MCP capabilities available to that user. Existing roles and permissions continue to apply, so connecting an AI client does not bypass your established access controls.

Flexible AI, on your terms

Your AI, your models, your infrastructure

OpenProject does not lock you into a specific AI provider or model. The MCP Server is built on the open Model Context Protocol (MCP), so you can connect compatible commercial AI assistants, open-source clients, or self-hosted language models.

You choose which AI technology to use, where it runs, and which services receive project data. This gives organizations the freedom to match their AI setup to requirements for data protection, security, compliance, and infrastructure.

Illustration showing flexible connections between AI models and infrastructure.

Get started with responsible AI

Connect your AI with the OpenProject MCP Server

The MCP Server is the bridge between OpenProject and your AI tools. Configure the connection, then let teams work with project context from a compatible AI assistant or agent.

1. Enable MCP. Configure the OpenProject MCP Server for your organization.

2. Connect your AI. Choose an MCP-compatible client and model that meet your requirements.

3. Define how agents can work. Control authentication, permissions, available MCP capabilities, and where data is processed.

Get started with MCP

Frequently asked questions

OpenProject is building AI capabilities that help teams work more efficiently with their project and portfolio data. The OpenProject MCP Server already enables compatible AI assistants to interact with OpenProject. Further capabilities, including AI-assisted text creation, semantic search, reporting and project management support, are part of OpenProject’s ongoing AI development.

AI can help you find and understand project information faster, summarize project data, support reporting and reduce repetitive work. OpenProject is also exploring AI assistance for text creation, semantic search, project management guidance and portfolio-level insights.

The OpenProject MCP Server is already available and provides ead and writeEnd of the insertion access to OpenProject information for compatible AI tools. Additional capabilities described in the OpenProject roadmap, including AI text assistance, semantic search and further AI-supported project and portfolio management workflows, are being developed or planned.

The MCP Server implements the Model Context Protocol, letting AI assistants and other tools securely interact with project data like projects, work packages, comments, and relations.

Yes, provided the AI application supports the Model Context Protocol (MCP) and can connect to the OpenProject MCP Server. MCP provides a standardized way for compatible AI assistants to work with up-to-date OpenProject information.

No. OpenProject’s approach is designed to provide flexibility rather than lock organizations into one AI provider. The MCP Server uses an open standard, allowing organizations to choose compatible AI tools and models that fit their requirements.

OpenProject puts data sovereignty and organizational control at the center of its AI approach. The MCP Server uses OpenProject’s authentication mechanisms, and administrators control whether the integration is enabled and what tools are available. When connecting an external AI provider, how data is processed also depends on the AI service and deployment architecture you choose.

Partly. The OpenProject MCP Server lets connected AI assistants create and update work packages, add comments, and add, change or remove relations. Deleting work packages is not possible, and assistants only ever act with the permissions of the connected user. Support for further data such as time entries and meetings is in progress.

Yes. Administrators remain in control of whether the OpenProject MCP Server is available on their instance and can configure the integration according to their organization’s requirements. This allows organizations to decide how and whether AI integrations are used.

NOT SURE YET?

Try OpenProject free

Start free trial
Search area: