# Concord MCP for AI Agents MCP

> Concord provides your AI agent with full programmatic access to your self-hosted CI/CD workflow orchestration platform. It lets you manage organizations, view project structures, run new deployments, track active processes, and pull detailed execution logs—all from natural language commands.

## Overview
- **Category:** industry-titans
- **Price:** Free
- **Tags:** workflow-orchestration, ci-cd, process-automation, log-analysis, pipeline-management

## Description

Managing complex workflows usually means jumping between dashboards: checking the status dashboard, pulling logs into a text editor, cross-referencing project details in another tab. This MCP changes that. It connects your agent directly to your Concord instance, giving it visibility into every part of your CI/CD process.

You can ask your AI client to list all organizations and projects across your entire setup. Need to check a failing deployment? Simply ask the agent for the logs for a specific run, and it retrieves the failure context instantly. You don't have to manually copy IDs or navigate deep into menus; you just describe what you need—a running process status, project structure, or log output.

Whether you are troubleshooting an issue during an incident response or simply auditing your system's current state, this MCP centralizes that operational knowledge. When connected through Vinkius, it makes Concord a natural part of your existing AI toolset, allowing your agent to handle complex process management tasks without needing dedicated UI interaction.

## Tools

### list_projects
Gets a list of all projects within a specified organization.

### list_repositories
Retrieves a list of repositories that are configured for a given project.

### list_running_processes
Quickly lists all processes that are currently running in the workflow system.

### start_process
Triggers a new workflow execution run for a defined project or template.

### terminate_process
Stops an active workflow process execution immediately.

### get_process
Retrieves detailed status and metadata about a single, specific process run.

### get_process_log
Fetches the full text logs for any given workflow execution instance.

### get_project_details
Retrieves comprehensive details about a specific project within Concord.

### list_organizations
Gets an inventory list of every organization configured in your system.

### list_processes
Retrieves a full history listing of all past and present process executions.

## Prompt Examples

**Prompt:** 
```
Show me all active deployments that are currently running or have failed in the last 24 hours.
```

**Response:** 
```
Here's a summary of recent activity:

*   **Project Phoenix:** Running (Started 5 minutes ago). Status: In Progress. Last task: Build container image.
*   **Billing API:** Failed (Completed 3 hours ago). Error context: Connection Timeout at step 'Validate credentials'.
*   **Frontend UI:** Successful (Completed 1 day ago). Final status: Deployed to Staging.
```

**Prompt:** 
```
I need the full logs for the Billing API failure. What was wrong?
```

**Response:** 
```
Pulling detailed execution logs now... 

**[Log Snippet Start]**

`Task 'Validate credentials' failed: ECONNREFUSED 127.0.0.1:5432`. The service appears to be inaccessible or the port is closed.

**[Log Snippet End]**

The agent suggests checking network connectivity first.
```

**Prompt:** 
```
List all organizations and their projects so I can plan the next audit.
```

**Response:** 
```
Your Concord instance manages 4 distinct organizational units:

1.  **Engineering:** 24 active projects.
2.  **Data Platform:** 8 active projects (Focus: ETL workflows).
3.  **QA Automation:** 15 active projects.
4.  **Finance Systems:** 3 active projects.

Which organization do you want to investigate next?
```

## Capabilities

### Discovering Project Scope
Retrieve the full organizational structure, listing all configured organizations and their contained projects.

### Tracking Workflow Status
Get a live overview of every currently running workflow process and retrieve detailed metrics for any specific instance.

### Controlling Processes
Trigger new deployments or manually stop runaway processes using conversational commands.

### Diagnosing Failures
Pull the complete, historical execution logs for any process instance to find the exact failure context.

## Use Cases

### Investigating an Overnight Failure
A Release Manager notices that staging deployment failed. Instead of hunting through logs manually, they ask their agent to pull the logs for the specific process run ID, immediately identifying the database connection error and narrowing down the fix.

### System Audit Before Expansion
A Platform Team needs to know how many business units are running workflows. They use list_organizations to pull a definitive count of all existing organizations, which helps them scope out required resource allocations for expansion.

### Stopping a Bad Deployment
During testing, an engineer realizes a process is running with incorrect credentials and could cause damage. They immediately tell their agent to list_running_processes to find the job ID and then terminate_process before it finishes.

### Mapping Project Scope
A new team member needs to understand all deployment targets. They ask their agent to get_project_details for a main product line, which returns not just the project name but also every associated repository and its purpose.

## Benefits

- Stop context switching. Instead of opening the dashboard, navigating to 'Processes,' finding the ID, and then pulling logs into a separate window, you simply ask your agent for the details using the get_process_log tool.
- Audit quickly across large systems. Need to know what departments exist? Use list_organizations to pull an inventory list of every single organizational silo in minutes.
- Handle incidents without hands-on access. If a deployment is stuck, you can use list_running_processes to identify the runaway job and then terminate_process to shut it down immediately.
- Understand project dependencies instantly. By calling list_projects or get_project_details, your agent provides a clear map of which repositories are linked to which projects.
- Automate execution triggers. Instead of manually clicking 'Run' on a build job, you can use start_process and let your agent handle the workflow initiation step.

## How It Works

The bottom line is that you manage your entire CI/CD pipeline conversationally, without leaving your chat interface.

1. Add the Concord integration details—your instance URL and API Token—to your AI client's toolset.
2. Your agent uses the stored credentials to connect directly to your private CI/CD workflow platform.
3. You issue a natural language command (e.g., 'Show me the logs for project X'), and the agent executes the necessary action via Concord, returning the structured data or text output.

## Frequently Asked Questions

**How does Concord MCP help me track deployments?**
The agent provides a unified view of your entire CI/CD lifecycle. You can list all processes, see which ones are running right now, and get status updates on past runs without switching tabs or dashboards.

**Can I use Concord MCP to find out what repositories exist?**
Yes. If you tell the agent a specific project name, it can retrieve all associated repositories for that project. This is useful for auditing dependencies and understanding your code base scope.

**What if my deployment fails? Can Concord MCP help me debug?**
Absolutely. You ask the agent to pull the logs for a specific process run, and it retrieves the complete text output, letting you see the exact error context—like which connection failed or what line of code broke.

**Does Concord MCP only work with my self-hosted instance?**
Yes. This MCP is designed to connect specifically to your own private, self-hosted Concord deployment. It keeps all your workflow data securely within your environment.

**How do I start a new process using the Concord MCP?**
You simply ask your agent to trigger a run for a specific project. The agent handles the necessary API calls, starts the deployment sequence in Concord, and reports back when it begins.

**Is listing organizations part of the Concord MCP? **
Yes, this is one of its core capabilities. It gives you a high-level inventory of every organizational silo within your system, which is critical for large platform audits and scoping new work.