bitbucket-server-mcp
Enables to interact with Bitbucket Server repositories, pull requests, and code reviews, including file browsing, PR management, and review actions through natural language.
README
<div align="center">
</div>
<div align="center"> <img src="docs/demo.gif" alt="Demo: listing pending PR reviews with Claude Code" width="800"> </div>
Quickstart
Or via Claude Code:
claude mcp add bitbucket \
-e BITBUCKET_URL=https://your-bitbucket-server.com \
-e BITBUCKET_TOKEN=your-token \
-- npx -y @upendra_sengar/bitbucket-server-mcp
Requirements
- Bitbucket Server / Data Center 7.21+ (the E2E suite covers 7.21, 8.5, 8.9, 8.19, 9.4 and 10.2).
- One of:
Installation
<details> <summary><strong>VS Code</strong></summary>
Add to your workspace .vscode/mcp.json:
{
"servers": {
"bitbucket": {
"command": "npx",
"args": ["-y", "@upendra_sengar/bitbucket-server-mcp"],
"env": {
"BITBUCKET_URL": "https://your-bitbucket-server.com",
"BITBUCKET_TOKEN": "your-access-token"
}
}
}
}
</details>
<details> <summary><strong>Claude Desktop, Cursor, Windsurf, JetBrains, Neovim</strong></summary>
These clients all use the same mcpServers format. Add the JSON below to the config file for your client:
| Client | Config file |
|---|---|
| Claude Desktop | ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows) |
| Cursor | ~/.cursor/mcp.json (one-click install) |
| Windsurf | ~/.codeium/windsurf/mcp_config.json |
| JetBrains | Settings > Tools > AI Assistant > MCP, or ~/.junie/mcp/mcp.json |
| Neovim | mcphub.nvim mcpservers.json |
{
"mcpServers": {
"bitbucket": {
"command": "npx",
"args": ["-y", "@upendra_sengar/bitbucket-server-mcp"],
"env": {
"BITBUCKET_URL": "https://your-bitbucket-server.com",
"BITBUCKET_TOKEN": "your-access-token"
}
}
}
}
</details>
<details> <summary><strong>Zed</strong></summary>
Add to your Zed settings (~/.config/zed/settings.json on Linux, ~/Library/Application Support/Zed/settings.json on macOS):
{
"context_servers": {
"bitbucket": {
"source": "custom",
"command": "npx",
"args": ["-y", "@upendra_sengar/bitbucket-server-mcp"],
"env": {
"BITBUCKET_URL": "https://your-bitbucket-server.com",
"BITBUCKET_TOKEN": "your-access-token"
}
}
}
}
</details>
<details> <summary><strong>Docker</strong></summary>
For environments without Node.js:
{
"mcpServers": {
"bitbucket": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "BITBUCKET_URL=https://your-bitbucket-server.com",
"-e", "BITBUCKET_TOKEN=your-access-token",
"ghcr.io/upendrasengar/bitbucket-server-mcp"
]
}
}
}
Or build locally: docker build -t bitbucket-mcp .
</details>
[!TIP] Bun users can replace
npx -ywithbunxin any of the configs above.
Tools
Repositories
| Tool | Description |
|---|---|
list_projects |
List all accessible Bitbucket projects |
list_repositories |
List repositories in a project |
browse_repository |
Browse files and directories |
get_file_content |
Read file contents with pagination |
upload_attachment |
Upload a local file and get a markdown reference for PR comments |
edit_file |
Edit a file by committing a new version via the REST API |
get_server_info |
Get Bitbucket Server version and properties |
get_file_blame |
Get line-by-line blame/history for a file |
create_repository |
Create a new repository |
delete_repository |
Delete a repository (irreversible) |
Pull Requests
| Tool | Description |
|---|---|
create_pull_request |
Create a PR, including cross-repo from forks (sourceProject/sourceRepository) and drafts (draft: true). Auto-fetches default reviewers unless includeDefaultReviewers: false. |
get_pull_request |
Get PR details |
update_pull_request |
Safely update title, description, or reviewers (read-modify-write, preserves fields not explicitly changed) |
merge_pull_request |
Merge a PR with optional strategy (no-ff, ff, ff-only, squash, rebase-no-ff, rebase-ff-only, squash-ff-only) |
decline_pull_request |
Decline a PR |
list_pull_requests |
List PRs with filtering by state, author, direction |
list_dashboard_pull_requests |
List PRs across all repos for the authenticated user, filtered by role (AUTHOR/REVIEWER/PARTICIPANT), state, and review status |
get_pull_request_activity |
Get PR activity timeline, filtered by type (all, reviews, comments) |
get_diff |
Get PR diff with per-file truncation support. Use stat: true for a lightweight summary of changed files instead of the full diff. |
get_pull_request_commits |
List commits in a pull request |
get_commit_pull_requests |
List pull requests containing a specific commit |
Code Review
| Tool | Description |
|---|---|
manage_comment |
Unified create/edit/delete for PR comments. Supports inline anchoring (filePath/line/lineType), draft state (state: PENDING), task creation (severity: BLOCKER), threaded replies (parentId), and resolve/unresolve (state: RESOLVED/OPEN). |
manage_review |
Unified approve/unapprove/publish. Publish transitions all PENDING comments to visible and optionally sets participantStatus (APPROVED/NEEDS_WORK). |
Branches & Commits
| Tool | Description |
|---|---|
list_branches |
List branches with default branch detection |
list_commits |
Browse commit history with branch and author filtering |
manage_branches |
Create or delete branches (safety check prevents deleting default branch) |
get_commit |
Get details of a specific commit by its ID |
compare_refs |
List commits reachable from one ref but not another |
list_tags |
List tags in a repository |
manage_tags |
Create or delete tags |
get_tag |
Get details of a specific tag by its name |
Search & Insights
| Tool | Description |
|---|---|
search |
Search code and files across repositories |
get_code_insights |
Fetch Code Insights reports (SonarQube, security scans) and annotations |
get_build_status |
Get CI build status (state, name, URL) by commit ID or PR. When using prId, resolves the latest commit automatically. |
Forks
| Tool | Description |
|---|---|
list_forks |
List forks of a repository |
fork_repository |
Fork a repository into a target project |
Users
| Tool | Description |
|---|---|
get_user_profile |
Get a user profile by slug |
search_users |
Search users by filter query |
Labels
| Tool | Description |
|---|---|
list_labels |
List labels for a repository |
manage_labels |
Add or remove repository labels |
Webhooks
| Tool | Description |
|---|---|
list_webhooks |
List webhooks for a repository |
manage_webhooks |
Create, update, or delete webhooks |
Commit Comments
| Tool | Description |
|---|---|
list_commit_comments |
List comments on a specific commit |
manage_commit_comments |
Create, edit, or delete comments on a commit |
Repository Settings
| Tool | Description |
|---|---|
list_default_reviewer_conditions |
List default reviewer conditions |
list_branch_restrictions |
List branch permission restrictions |
list_repository_hooks |
List repository hooks and their status |
manage_repository_hooks |
Enable, disable, or configure repository hooks |
list_merge_checks |
List merge check configurations |
manage_merge_checks |
Configure merge check settings |
list_reviewer_groups |
List reviewer groups |
manage_reviewer_groups |
Create or delete reviewer groups |
list_secret_scanning_rules |
List secret scanning allowlist rules (8.5+) |
Keys
| Tool | Description |
|---|---|
list_ssh_keys |
List SSH keys for the authenticated user |
manage_ssh_keys |
Add or delete SSH keys |
list_gpg_keys |
List GPG keys for the authenticated user |
manage_gpg_keys |
Add or delete GPG keys |
Prompts
| Prompt | Description |
|---|---|
review-pr |
Step-by-step workflow for reviewing a PR: fetch details, read diff, check CI, create draft comments, and publish the review. Invoke via /bitbucket:review-pr in Claude Code. No arguments needed; the LLM asks for the PR interactively. You can add context, e.g. /bitbucket:review-pr PR #42 in my-repo. |
Resources
| Resource | URI | Description |
|---|---|---|
projects |
bitbucket://projects |
Cached list of all accessible projects (5 min TTL). Useful as ambient context without explicit tool calls. |
Configuration
Environment Variables
| Variable | Required | Description |
|---|---|---|
BITBUCKET_URL |
Yes | Base URL of your Bitbucket Server instance |
BITBUCKET_TOKEN |
Yes* | Personal access token |
BITBUCKET_USERNAME |
Yes* | Username for basic auth |
BITBUCKET_PASSWORD |
Yes* | Password for basic auth |
BITBUCKET_DEFAULT_PROJECT |
No | Default project key when not specified in tool calls |
BITBUCKET_READ_ONLY |
No | Set to true to disable all write operations |
BITBUCKET_CUSTOM_HEADERS |
No | Extra headers for all requests (Key=Value,Key2=Value2). Useful for Zero Trust tokens. |
BITBUCKET_DIFF_MAX_LINES_PER_FILE |
No | Max lines per file in diffs. Set to 0 for no limit. |
BITBUCKET_CACHE_TTL |
No | Cache duration in seconds (default: 300). Set to 0 to disable caching. |
BITBUCKET_ENABLED_TOOLS |
No | Comma-separated list of tool names to enable. If not set, all tools are available. |
BITBUCKET_STARTUP_HEALTHCHECK |
No | Set to true to run a connectivity check against Bitbucket on startup (default: false). |
*Either BITBUCKET_TOKEN or both BITBUCKET_USERNAME and BITBUCKET_PASSWORD are required.
Read-Only Mode
Set BITBUCKET_READ_ONLY=true to restrict the server to read-only operations. All create_*, update_*, delete_*, manage_*, merge_*, decline_*, fork_*, upload_*, and edit_* tools are disabled.
Tool Filtering
Set BITBUCKET_ENABLED_TOOLS to load only specific tools, reducing context window usage:
BITBUCKET_ENABLED_TOOLS=get_pull_request,get_diff,manage_comment,manage_review
Startup Healthcheck
When BITBUCKET_STARTUP_HEALTHCHECK=true, the server probes
/rest/api/1.0/application-properties on startup and logs whether Bitbucket is
reachable. If not, it logs a diagnostic line referencing the relevant env vars
(BITBUCKET_URL, BITBUCKET_TOKEN, HTTPS_PROXY, NODE_EXTRA_CA_CERTS)
without blocking the server. Useful for debugging connectivity issues before
the first tool call fails with a generic error.
BITBUCKET_STARTUP_HEALTHCHECK=true
Response Curation
Read tools return compact responses by default, keeping only the fields an AI assistant typically needs. Every read tool accepts a fields parameter to customize:
- Omit
fields: returns a curated summary (e.g. PR id, title, state, author, branches, reviewers, task count) fields: "*all": returns the complete raw Bitbucket API responsefields: "id,title,author.user.name": returns exactly those fields (dot notation for nested paths)
Caching
The server caches frequently accessed data in memory (project lists, repository metadata, default reviewers) to reduce API calls. The cache uses LRU eviction (max 500 entries) so memory stays bounded, and write operations automatically invalidate related entries.
By default, cached entries expire after 5 minutes. Configure with BITBUCKET_CACHE_TTL (in seconds), or set to 0 to disable caching entirely.
Contributing
See CONTRIBUTING.md for development setup, architecture overview, and how to add new tools.
License
Apache 2.0
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.