A Model Context Protocol (MCP) server for Linear, written in Go. This server provides tools for interacting with the Linear API through the MCP protocol.
- Create, update, and search Linear issues
- Get issues assigned to a user
- Add comments to issues and reply to existing comments
- URL-aware comment operations - paste Linear comment URLs directly, no manual ID extraction needed
- Retrieve team information
- Rate-limited API requests to respect Linear's API limits
- Go 1.23 or higher
- Linear API key
Pre-built binaries are available for Linux, macOS, and Windows on the GitHub Releases page.
- Download the appropriate binary for your platform
- Make it executable (Linux/macOS):
chmod +x linear-mcp-go-*
- Run the binary as described in the Usage section
# Download linux binary for the latest release
RELEASE=$(curl -s https://api.github.com/repos/geropl/linear-mcp-go/releases/latest)
DOWNLOAD_URL=$(echo $RELEASE | jq -r '.assets[] | select(.name | contains("linux")) | .browser_download_url')
curl -L -o ./linear-mcp-go $DOWNLOAD_URL
chmod +x ./linear-mcp-go
# Setup the mcp server (.gitpod.yml, dotfiles repo, etc.)
./linear-mcp-go setup --tool=cline
To check the version of the Linear MCP server:
./linear-mcp-go version
This will display the version, git commit, and build date information.
- Set your Linear API key as an environment variable:
export LINEAR_API_KEY=your_linear_api_key
- Run the server:
# Run in read-only mode (default)
./linear-mcp-go serve
# Run with write access enabled
./linear-mcp-go serve --write-access
The server will start and listen for MCP requests on stdin/stdout.
The setup
command automates the installation and configuration process for various AI assistants:
# Set your Linear API key as an environment variable
# Only exception: Ona does not require this for setup!
export LINEAR_API_KEY=your_linear_api_key
# Set up for Cline (default)
./linear-mcp-go setup
# Set up with write access enabled
./linear-mcp-go setup --write-access
# Set up with auto-approval for read-only tools
./linear-mcp-go setup --auto-approve=allow-read-only
# Set up with specific tools auto-approved
./linear-mcp-go setup --auto-approve=linear_get_issue,linear_search_issues
# Set up with write access and auto-approval for read-only tools
./linear-mcp-go setup --write-access --auto-approve=allow-read-only
# Set up for a different tool (only "cline" supported for now)
./linear-mcp-go setup --tool=cline
This command:
- Checks if the Linear MCP binary is already installed
- Copies the current binary to the installation directory if needed
- Configures the AI assistant to use the Linear MCP server
- Sets up auto-approval for specified tools if requested
The --auto-approve
flag can be used to specify which tools should be auto-approved in the Cline configuration:
--auto-approve=allow-read-only
: Auto-approves all read-only tools (linear_search_issues
,linear_get_user_issues
,linear_get_issue
,linear_get_teams
)--auto-approve=tool1,tool2,...
: Auto-approves the specified comma-separated list of tools
Currently supported AI assistants:
- Cline (VSCode extension)
By default, the server runs in read-only mode, which means the following tools are disabled:
linear_create_issue
linear_update_issue
linear_add_comment
linear_reply_to_comment
linear_update_issue_comment
To enable these tools, use the --write-access=true
flag.
Creates a new Linear issue with specified details. Supports creating parent-child relationships (sub-issues) and assigning labels.
Parameters:
title
(required): Issue titleteam
(required): Team identifier (key, UUID or name)description
: Issue descriptionpriority
: Priority. Accepts: 0/'no priority', 1/'urgent', 2/'high', 3/'medium', 4/'low'status
: Issue statusmakeSubissueOf
: Create a sub-issue by specifying the parent issue ID or identifier (e.g., 'TEAM-123'). This establishes a parent-child relationship in Linear.labels
: Optional comma-separated list of label IDs or names to assignproject
: Optional project identifier (ID, name, or slug) to assign the issue to
Example: Creating a sub-issue
{
"title": "Implement login form validation",
"team": "ENG",
"makeSubissueOf": "ENG-42",
"description": "Add client-side validation for the login form"
}
Updates an existing Linear issue's properties.
Parameters:
id
(required): Issue IDtitle
: New titledescription
: New descriptionpriority
: Priority. Accepts: 0/'no priority', 1/'urgent', 2/'high', 3/'medium', 4/'low'status
: New status
Searches Linear issues using flexible criteria.
Parameters:
query
: Optional text to search in title and descriptionteamId
: Filter by team IDstatus
: Filter by status name (e.g., 'In Progress', 'Done')assigneeId
: Filter by assignee's user IDlabels
: Filter by label names (comma-separated)priority
: Priority. Accepts: 0/'no priority', 1/'urgent', 2/'high', 3/'medium', 4/'low'estimate
: Filter by estimate pointsincludeArchived
: Include archived issues in results (default: false)limit
: Max results to return (default: 10)
Retrieves issues assigned to a specific user or the authenticated user.
Parameters:
userId
: Optional user ID. If not provided, returns authenticated user's issuesincludeArchived
: Include archived issues in resultslimit
: Maximum number of issues to return (default: 50)
Retrieves a single Linear issue by its ID.
Parameters:
issueId
(required): ID of the issue to retrieve
Adds a comment to an existing Linear issue. Supports replying to existing comments by passing a comment identifier in the thread
parameter.
Parameters:
issue
(required): ID or identifier (e.g., 'TEAM-123') of the issue to comment onbody
(required): Comment text in markdown formatthread
: Optional comment identifier to reply to. Accepts: full Linear comment URL, UUID, shorthand (comment-abc123), or hash (abc123). Creates a threaded reply instead of a top-level comment.createAsUser
: Optional custom username to show for the comment
URL Support: You can pass a full Linear comment URL (e.g., https://linear.app/.../issue/TEST-10/...#comment-abc123
) directly to the thread
parameter. The tool automatically resolves URLs to UUIDs before calling the API.
Convenience tool for replying to an existing comment. Automatically resolves the issue from the comment, so you only need to provide the comment identifier and reply text.
Parameters:
thread
(required): Comment to reply to. Accepts: full Linear comment URL, UUID, shorthand (comment-abc123), or hash (abc123)body
(required): Reply text in markdown formatcreateAsUser
: Optional custom username to show for the reply
Why use this tool? When you have a comment URL or ID and want to reply, this tool is simpler than linear_add_comment
because you don't need to specify the issue separately. The tool automatically looks up the issue from the comment.
Retrieves comments for a Linear issue with support for pagination and thread navigation.
Parameters:
issue
(required): ID or identifier (e.g., 'TEAM-123') of the issue to retrieve comments forthread
: Optional UUID of a parent comment to retrieve its replies. If not provided, returns top-level commentslimit
: Maximum number of comments to return (default: 10)after
: Cursor for pagination, to get comments after this point
Use Cases:
- View all comments on an issue
- Navigate comment threads by passing a comment UUID in the
thread
parameter - Get comment UUIDs for replying (though with URL support in
linear_add_comment
, this is less necessary)
Updates an existing comment on a Linear issue.
Parameters:
comment
(required): Comment identifier to update. Accepts: full Linear comment URL, UUID, shorthand (comment-abc123), or hash (abc123)body
(required): New comment text in markdown format
URL Support: Like other comment tools, this accepts full Linear comment URLs and automatically resolves them to UUIDs.
Retrieves Linear teams with an optional name filter.
Parameters:
name
: Optional team name filter. Returns teams whose names contain this string.
Tests are implemented using go-vcr
, and executed against https://linear.app/linear-mcp-go-test.
Using the existing recordings (cassettes):
go test -v ./...
Requires TEST_LINEAR_API_KEY
to be set for the test workspace.
go test -v -record=true ./...
This will update all tests that don't alter remote state.
go test -v -recordWrites=true ./...
This will re-run all tests, including some that might alter the outcome of other tests cases, which might require further manual work to adjust.
go test -v -golden=true ./...
Updates all .golden fields.
The project uses GitHub Actions for automated testing and releases. The version is managed through the ServerVersion
constant in pkg/server/server.go
.
- All pushes to the main branch and pull requests are automatically tested
- When a tag matching the pattern
v*
(e.g.,v1.0.0
) is pushed, a new release is automatically created - Binaries for Linux, macOS, and Windows are built and attached to the release with build-time information (git commit and build date)
Important: Version tags should only be created against the main
branch after all changes have been merged.
-
Update the version: Modify the
ServerVersion
constant inpkg/server/server.go
// ServerVersion is the version of the MCP server ServerVersion = "1.13.0"
-
Create a PR: Submit the version update as a pull request to ensure it goes through review and testing
-
Merge to main: Once the PR is approved and merged to the main branch
-
Create and push the release tag:
# Ensure you're on the latest main branch git checkout main git pull origin main # Create and push the tag (must match the version in server.go) git tag v1.13.0 git push origin v1.13.0
-
Automated release: The GitHub Actions workflow will automatically:
- Build binaries for all platforms with proper version information
- Create a GitHub release with the tag
- Attach the compiled binaries to the release
The version
command displays:
- Version: Read from
ServerVersion
constant inpkg/server/server.go
- Git commit: Injected at build time from the current commit hash
- Build date: Injected at build time with the current timestamp
For development builds, git commit and build date will show "unknown".
MIT