Gcore MCP Server
Gcore Cloud API management tools for LLM-based cloud infrastructure automation.
Data last scanned today · Reviewed today
Overview
The Gcore MCP Server enables direct interaction with Gcore Cloud services through AI assistants. It provides a structured interface for managing cloud instances, bare metal clusters, and management resources by leveraging the Gcore API. Users can configure specific tool subsets to optimize performance and prevent token overhead in LLM context windows. Installation is streamlined through uvx, allowing for temporary execution without persistent local deployment. The server supports detailed environment configurations, including project and region IDs, to facilitate precise cloud resource control within development environments.
Our verdict
With a security score of 43/100, this server is currently classified as high risk. Because it is a community-maintained project rather than an official vendor release, it has not undergone formal editorial review, and both its authentication methods and read-only support remain unverified. It is suitable for experimental cloud management tasks, but users should exercise caution given the potential for API-level actions.
- Zero direct dependencies reduce the supply-chain attack surface.
- Active development evidenced by daily commit frequency.
- Apache-2.0 open source license allows for transparent auditing.
- Five distinct contributors indicate a moderate level of community oversight.
- Lacks verification of security-critical read-only modes.
- Community-maintained status lacks official vendor oversight and endorsement.
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›What is the recommended way to configure tools to avoid overwhelming the LLM?
For most clients, it is recommended to specify only the necessary resources in the GCORE_TOOLS environment variable to ensure optimal performance and clarity.
›How does tool loading differ when using Claude Code?
Claude Code uses a deferred loading mechanism where tool schemas are only fetched on demand, allowing you to safely enable all tools via GCORE_TOOLS=* without bloating the context window.
›Does the HTTP transport provide built-in authentication for clients?
No, the HTTP transport does not authenticate clients. Anyone who can reach the listener can use the server's GCORE_API_KEY, so it should be bound to loopback and kept off untrusted networks.
›How can I customize which tools are available?
You can use the GCORE_TOOLS environment variable to define toolsets or use wildcard patterns to match specific tools from the Gcore SDK.
›Is the SSE transport supported?
No, the legacy SSE transport is not supported, and setting GCORE_TRANSPORT=sse will cause the server to refuse to start.
›What happens if I don't set a GCORE_TOOLS configuration?
The server defaults to the 'management' and 'instances' toolsets when using HTTP transport, and only the 'management' toolset when using stdio transport.
Alternatives
Badge
Maintain this server? Add the live badge to your README.