# OpenReplay MCP

> OpenReplay helps you debug web applications by letting your AI agent inspect recorded user sessions, technical events, and project data conversationally. Stop digging through logs; just ask your agent what went wrong for a specific user or feature.

## Overview
- **Category:** developer-tools
- **Price:** Free
- **Tags:** session-replay, debugging, user-experience, error-tracking, web-analytics

## Description

Your agent connects directly to OpenReplay so you can analyze user behavior without leaving the chat interface. You tell it what's broken—maybe a button isn't working on mobile, or an error pops up only sometimes—and your agent handles the heavy lifting. It finds the exact session recording and pulls out all the technical details you need. Whether you’re checking console logs, tracking specific clicks, or verifying if a user even exists in the project, it gets you to the source of truth fast. If your current AI client supports Vinkius, this MCP brings that deep debugging context right into your workflow. This means no more manual searches across different dashboards; you just talk to your agent and get actionable technical data.

## Tools

### get_session
Retrieves basic metadata for one specific session recording.

### list_projects
Gets a list of all projects connected to your account, helping you narrow down the scope.

### list_session_events
Pulls every technical action and user event recorded within an entire session.

### list_sessions
Retrieves a list of sessions for a specific project, giving you filters by User ID or page.

### search_users
Finds and verifies the existence of users within your projects using emails or IDs.

## Prompt Examples

**Prompt:** 
```
List all projects in my OpenReplay account.
```

**Response:** 
```
I've found 3 projects: 'Main Web App' (ID: 12), 'Mobile Dashboard' (ID: 45), and 'Staging Environment' (ID: 8). Which one would you like to explore?
```

**Prompt:** 
```
Find the last 5 sessions for user 'john@example.com' in project 12.
```

**Response:** 
```
Searching for the user... Found User ID 'user_99'. Here are the 5 most recent sessions for John, including a 12-minute session from today at 10:30 AM. Would you like to see the events for that session?
```

**Prompt:** 
```
Show me any console errors or clicks for session ID 'sess-abc-123'.
```

**Response:** 
```
Inspecting session 'sess-abc-123'... I found a 'Uncaught TypeError' at 02:15 and a series of rapid clicks on the 'Submit' button immediately following the error. It looks like a validation issue.
```

## Capabilities

### Identify projects
Lists all web application environments or projects linked to your OpenReplay account.

### Find user sessions
Retrieves a list of recorded user sessions for a specific project, allowing you to filter results by User ID.

### Search users
Locates and identifies specific users within your projects using their email addresses or unique IDs.

### Get session details
Pulls the metadata for a single, targeted user session recording.

### Retrieve technical events
Fetches all recorded technical and user actions, like clicks or console logs, that happened during a specific session.

## Use Cases

### A customer reports a form submission error on the mobile app.
Instead of asking the customer for screenshots, you prompt your agent using 'list_sessions' and filter by their User ID. You then use 'list_session_events' to pull all technical data from that specific session, pinpointing whether the failure was a JavaScript error or an unexpected API response.

### A developer needs to reproduce a rare console bug.
The team knows the user ID but not the session. They first use 'list_projects' to confirm the correct environment, then ask for sessions using 'list_sessions'. Once they find the timeframe, they get the metadata via 'get_session', and finally retrieve all technical events to see the exact sequence of actions that triggered the bug.

### The product team needs to analyze feature adoption.
They ask their agent to run a search using 'search_users' for all internal testers. They then use this list of IDs with 'list_sessions' to see if the new dashboard component is being used at all, and what path users take before they drop off.

### The support team needs to verify a user account.
Given just an email address from a ticket, you use 'search_users' to confirm if the person exists in your project. This saves time and lets you immediately pivot to checking recent sessions for that verified ID.

## Benefits

- You stop manually searching through session lists. Your agent uses 'list_sessions' to quickly pull up the right records, even if you only know a date range or a User ID.
- 'search_users' lets you correlate support tickets directly with recorded behavior. You don't have to copy an email and paste it into three different systems just to confirm who complained.
- When you need technical proof, the 'list_session_events' tool fetches every click, input, and console log for a session. This gives developers immediate context they can't get from simple error messages.
- You can immediately narrow down your search scope using 'get_session'. Once you have a specific session ID, your agent pulls the metadata so you know exactly what that recording is about before diving into the events.
- Product teams gain deep insights by asking to list projects first. This helps them understand which application environment—main or staging—they need to analyze user flow data for.

## How It Works

The bottom line is you talk to your agent, and it does the repetitive work of navigating OpenReplay's complex session records so you don't have to.

1. Subscribe to this MCP and enter your OpenReplay API Key.
2. Your AI client uses the provided credentials to connect directly to the OpenReplay platform.
3. You prompt your agent with a natural language request, like 'Find the last 5 sessions for John in Project X,' and it returns the necessary data.

## Frequently Asked Questions

**How do I use OpenReplay MCP to find sessions for a specific user?**
You first need to confirm the project using 'list_projects'. Then, prompt your agent to use 'list_sessions', making sure you include the User ID and the correct Project ID in your request.

**Does OpenReplay MCP help me find technical bugs?**
Yes. The key is the 'list_session_events' tool. It retrieves every technical action, including console logs, for a session, so you can pinpoint exactly what went wrong.

**What if I don't know the user ID?**
You can use 'search_users'. This tool lets your agent find users based only on their email address or name within a specified project, giving you the needed ID to continue your debugging.

**Which tools in OpenReplay MCP are best for product managers?**
For analyzing user journeys and feature usage, 'list_sessions' combined with 'search_users' is most helpful. It lets you track groups of users to see how they interact over time.

**Can OpenReplay MCP tell me the project name?**
Yes. You use the 'list_projects' tool, which provides a list of all associated projects and their unique IDs, helping you scope your investigation correctly.