X Search MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| XAI_API_KEY | Yes | Your xAI API key for searching X (Twitter) and web via Grok. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| search_xA | Search X (Twitter) for real-time posts, discussions, and trends on any topic. Powered by Grok's x_search, returns summarized results with original quotes from users. Supports filtering by date range and specific X handles. |
| search_web_via_grokA | Search the web for real-time information via Grok. Returns summarized results from web pages. Useful for current events, documentation, news, and general knowledge. |
| search_combinedB | Search both X (Twitter) and the web simultaneously for comprehensive results. Grok will automatically decide which source to use or combine both. Best for getting a complete picture of a topic from social media AND web sources. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
search_x and search_web_via_grok clearly target distinct sources, but search_combined is a superset of both, so an agent may reasonably wonder why it would ever pick the narrower tools. The descriptions do clarify intent, keeping them mostly distinguishable.
All three share the search_ prefix, which is good, but the suffixes vary in style (search_x, search_web_via_grok, search_combined) rather than following one predictable pattern. Still readable and mostly consistent.
Three tools is on the thin side for a search server, and one of them (search_combined) arguably subsumes the other two, making the surface feel slightly redundant rather than lean.
The core search domain is covered across X, web, and combined sources, but finer operations like thread retrieval, user/timeline lookup, or pagination are absent. Minor gaps an agent can work around.