TestOps MCP Server
Enables AI assistants to interact with TestOps 5.25 for managing projects, test runs, test plans, and test cases via MCP tools.
README
TestOps MCP Server
Model Context Protocol server for TestOps 5.25
Installation
npm install @syn7xx/testops-mcp-server
Or use directly via npx:
npx @syn7xx/testops-mcp-server --url <testops-url> --token <api-token>
Short flags: -u / -t / -h (same as --url, --token, --help).
Environment variables: TESTOPS_URL, TESTOPS_TOKEN (used when flags are omitted).
Build
npm run build
Run
npm start -- --url <testops-url> --token <api-token>
Or with npx (if not installed globally):
npx @syn7xx/testops-mcp-server --url <testops-url> --token <api-token>
TLS verification (optional)
If TestOps is served with a self-signed certificate or HTTPS is intercepted by a corporate proxy and Node.js fails TLS validation, you can disable certificate verification for that Node process by setting:
export NODE_TLS_REJECT_UNAUTHORIZED=0
Then start the server as usual (npm start, npx, or your MCP command). In MCP configs that support server environment variables, set the same name next to TESTOPS_URL / TESTOPS_TOKEN (e.g. OpenCode environment, or your client’s env field).
Security: this disables TLS peer verification for all HTTPS calls from that process. Prefer installing the correct CA in the system trust store; use NODE_TLS_REJECT_UNAUTHORIZED=0 only when necessary and in a trusted environment.
MCP setup in AI tools
Configs differ by product:
| Product | Typical config file | Root key | Notes |
|---|---|---|---|
| Cursor, Windsurf, Kilo Code, Zed | mcp.json (paths vary) |
mcpServers |
command + args |
| Visual Studio Code | .vscode/mcp.json or user MCP config |
servers |
Official docs — not mcpServers |
| Claude Desktop | claude_desktop_config.json |
mcpServers |
Same idea as Cursor |
| OpenCode | opencode.jsonc (project or user config) |
mcp |
type: "local", command as array, env under environment |
Do not commit real tokens; use env vars or local-only config.
Cursor, Windsurf, Kilo Code, Zed
These use top-level mcpServers, each entry has command (string) and args (array).
Cursor — ~/.cursor/mcp.json or project .cursor/mcp.json:
{
"mcpServers": {
"testops": {
"command": "npx",
"args": [
"-y",
"@syn7xx/testops-mcp-server",
"--url",
"https://your-testops.com",
"--token",
"your-token"
]
}
}
}
Windsurf — ~/.config/windsurf/mcp.json or project .windsurf/mcp.json (same structure).
Kilo Code — ~/.kilocode/mcp.json or project .kilocode/mcp.json (same structure).
Zed — ~/.config/zed/mcp.json (same structure).
Visual Studio Code
VS Code stores MCP config in mcp.json with top-level servers, not mcpServers (see Add and manage MCP servers in VS Code and the MCP configuration reference).
Workspace file: .vscode/mcp.json, or run MCP: Open User Configuration for a global file.
{
"servers": {
"testops": {
"command": "npx",
"args": [
"-y",
"@syn7xx/testops-mcp-server",
"--url",
"https://your-testops.com",
"--token",
"your-token"
]
}
}
}
For sensitive values, prefer input variables or env files as described in the reference.
OpenCode
OpenCode does not use mcpServers. Define servers under mcp, set type: "local", pass the process as a command array (e.g. npx, -y, @syn7xx/testops-mcp-server, …), and optional env in environment (not env). See OpenCode MCP servers.
Example — opencode.jsonc in the project root or in the global OpenCode config:
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"testops": {
"type": "local",
"command": [
"npx",
"-y",
"@syn7xx/testops-mcp-server",
"--url",
"https://your-testops.com",
"--token",
"your-token"
],
"enabled": true
}
}
}
Using environment variables instead of flags:
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"testops": {
"type": "local",
"command": ["npx", "-y", "@syn7xx/testops-mcp-server"],
"environment": {
"TESTOPS_URL": "https://your-testops.com",
"TESTOPS_TOKEN": "your-token"
},
"enabled": true
}
}
}
Claude Desktop
macOS — ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"testops": {
"command": "npx",
"args": [
"@syn7xx/testops-mcp-server",
"--url",
"https://your-testops.com",
"--token",
"your-token"
]
}
}
}
On Windows, the config lives under %APPDATA%\Claude\ — see Anthropic’s Claude Desktop documentation for the exact path.
Global install (optional)
If you prefer a fixed binary path instead of npx:
npm install -g @syn7xx/testops-mcp-server
Then use "command": "testops-mcp-server" and "args": ["--url", "...", "--token", "..."] under mcpServers (Cursor, Claude, …) or servers (VS Code), or the command array in OpenCode.
Tools
Project
| Tool | Description |
|---|---|
project_list |
List projects with pagination |
project_find_by_name |
Find project by exact or partial name match |
project_get_by_id |
Get project by ID |
Launch (test run)
| Tool | Description |
|---|---|
launch_create |
Create a launch directly |
launch_stop |
Close/stop a launch (end the run) |
launch_get_statistic |
Run summary: counts by status + progress ready |
launch_list_test_results |
Flat paginated test results for a launch (sort can be a string or array for multiple sort params) |
Test Plan
| Tool | Description |
|---|---|
testplan_get |
Get test plan by ID |
testplan_get_test_cases |
Get test cases from test plan with pagination |
testplan_run |
Start a launch from a test plan |
testplan_sync |
Sync test plan with source |
Test Case
Routing is defined in each tool’s description in tools/list. For step actions and expected results as scenario JSON, use testcase_get_scenario or testcase_get_step (same behavior).
| Tool | Description |
|---|---|
testcase_get |
Get test case by ID |
testcase_get_detail |
Summary with flattened step strings and custom fields (not raw scenario JSON) |
testcase_get_scenario |
Scenario JSON: steps and expected results |
testcase_get_step |
Same as testcase_get_scenario (alias for LLM-friendly naming) |
testcase_update_step |
Update a step in scenario |
testcase_set_scenario |
Replace all steps in scenario |
testcase_get_custom_fields |
Get custom field values for a test case |
testcase_update_custom_fields |
Set custom field values on a test case |
testcase_search_by_aql |
Search test cases using AQL query |
testcase_list_in_tree |
List test cases in a project tree with pagination |
testcase_create |
Create a test case with optional steps, tags, custom fields, precondition/postcondition |
Project Custom Fields
| Tool | Description |
|---|---|
project_get_custom_fields |
Get custom field definitions for a project with pagination ({ items, total, page, size }) |
project_get_custom_field_values |
Get available values for a custom field with pagination ({ items, total, page, size }) |
Project Structure
src/
├── shared/openapi/ # HTTP DTOs (common-dto, project-dto, test-plan-dto, test-case-dto, launch-dto, launch-test-result-dto)
├── shared/ # Utilities, API client, pagination
├── domain/
│ ├── test-case/ # Business logic (service, scenario, custom-fields, create)
│ └── ... # Other domains (project, launch, test-plan)
├── presentation/
│ ├── tool-utils.ts # handleResult helper for tools
│ ├── test-case/
│ │ └── tools/ # MCP tools (register, testcase-read, testcase-write, testcase-list, project-cf, schemas)
│ └── ... # Other presentations (project, launch, test-plan)
├── index.ts # Entry point
└── server.ts # MCP server configuration
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。