Use a job-first starter stack
Do not install a generic bundle of 'must-have' servers. Pick the one external system that repeatedly blocks your Cursor workflow. Docs/context is a common first candidate; browser/testing, repository, database, and design follow only when the built-in agent or a trusted CLI cannot satisfy the same task clearly.
How the live Cursor shortlist is computed
The data table includes active developer-workflow entries that declare Cursor compatibility and clear the displayed star floor. Duplicate repositories are collapsed, then TrustScore and adoption order the eligible set. The formula is repeatable and the pool size is printed below the table. A rank is a research starting point, not a safety or workflow guarantee.
| # | Server | TrustScore | Stars | Clients |
|---|---|---|---|---|
| 1 | io.github.D4Vinci/ScraplingScrapling MCP Server | 93 | 73,830 | cursor · derived |
| 2 | io.github.microsoft/playwright-mcpPlaywright Tools for MCP | 87 | 36,128 | cursor · derived |
| 3 | io.github.homeassistant-ai/ha-mcpComprehensive Model Context Protocol server for managing Home Assistant through | 87 | 4,381 | cursor · derived |
| 4 | io.github.bytebase/dbhubDBHub | 87 | 3,365 | cursor · derived |
| 5 | io.github.ChromeDevTools/chrome-devtools-mcpChrome DevTools MCP | 86 | 48,873 | cursor · derived |
| 6 | io.github.usekaneo/kaneoOfficial MCP server for Kaneo: manage tasks, projects, and labels from Claude an | 86 | 8,047 | cursor · derived |
| 7 | io.github.Ataraxy-Labs/semEntity-level code intelligence: semantic diff, impact analysis, blame, and conte | 86 | 3,302 | cursor · derived |
| 8 | io.github.atilaahmettaner/tradingview-mcpReal-time market data, screeners, technical analysis & backtesting for stocks, c | 85 | 3,937 | cursor · derived |
| 9 | io.github.zereight/gitlab-mcpGitLab MCP server for projects, merge requests, issues, pipelines, wiki, release | 85 | 1,886 | cursor · derived |
| 10 | io.github.github/github-mcp-serverGitHub | 84 | 32,250 | cursor · derived |
10 of 2077 actively-maintained servers in these categories cleared the 500-star adoption floor, ordered by TrustScore. Regenerated daily from public GitHub, npm and registry signals — see our editorial policy for the full weighting. ✓ means we installed it in a sandbox and confirmed it starts.
Separate derived compatibility from tested behavior
In the frozen snapshot, two catalog samples have stored sandbox startup/tool-list evidence and 2,398 Cursor rows are derived from transport metadata. This environment's Cursor CLI reports that no Cursor IDE is installed, so MCP Radar did not exercise discovery or a real tool invocation in the current stable IDE for this batch. Treat every selected row as requiring a local smoke test.
Configure and verify in Cursor
Put project-specific definitions in `.cursor/mcp.json` or personal global definitions in `~/.cursor/mcp.json`. Cursor supports stdio, SSE, and Streamable HTTP and documents MCP Logs in the Output panel. Enable one server, inspect its tools, run one non-destructive request, and verify the result and permission scope before enabling automatic execution.
Run the redundancy and exclusion tests
Compare the same task through Cursor's built-in capability or terminal command. Exclude the server if it merely wraps a familiar CLI without better schemas or controls, requests broad credentials, duplicates another server, has unclear ownership, or fails the current IDE smoke test. Remove unused servers to reduce tool clutter and attack surface.