Skip to main content
Glama
joaovfnandes

X Search MCP

by joaovfnandes
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Local

X Search MCP

Este MCP usa o CUA Driver para controlar o Microsoft Edge e pesquisar no X.com usando o mesmo perfil que já está aberto e autenticado.

Ferramentas

  • list_cua_tools: lista as ferramentas expostas pelo CUA conectado.

  • search_x: abre uma busca no X e retorna o texto atual da página.

  • get_latest_profile_posts: retorna os posts mais recentes visíveis de um perfil.

  • get_profile_info: retorna o texto acessível com as informações visíveis de um perfil.

  • search_hashtag_top: pesquisa uma hashtag e retorna os resultados mais relevantes.

  • search_hashtag_latest: pesquisa uma hashtag e retorna os posts mais recentes.

  • get_post_thread: abre um post e retorna o contexto visível da conversa.

search_x aceita operadores do X, por exemplo:

  • from:OpenAI model;

  • "frase exata";

  • since:2026-07-01 until:2026-07-15;

  • lang:pt filter:links.

O parâmetro live=true abre a aba de posts mais recentes. Ele apenas controla o filtro da busca; não é usado como mecanismo de atualização.

Related MCP server: Xee-mcp

Como funciona

  1. Localiza a janela aberta do Edge com list_windows.

  2. Navega pela barra de endereços usando get_window_state, click, set_value e press_key.

  3. Lê o conteúdo atual com page(action="get_text").

  4. Confirma que a URL e o texto retornado pertencem à consulta atual antes de responder.

Essa validação evita devolver silenciosamente o conteúdo de uma busca anterior quando a árvore de acessibilidade do X fica defasada.

Instalação

npm install
npm run build

Variáveis necessárias para o CUA via stdio:

CUA_TRANSPORT=stdio
CUA_COMMAND=C:\Users\User\AppData\Local\Programs\Cua\cua-driver\bin\cua-driver.exe
CUA_ARGS_JSON=["mcp"]

Para CUA via HTTP:

CUA_TRANSPORT=http
CUA_URL=http://localhost:3000/mcp

Teste de regressão

Com o Edge aberto e autenticado no X:

npm run test:sequential

O teste executa quatro pesquisas consecutivas e uma leitura de perfil na mesma sessão. Ele falha se alguma resposta contiver conteúdo defasado ou se a timeline do perfil ainda estiver carregando.

Para testar somente a timeline de um perfil:

npm run test:profile

Requisitos e limites

  • O Edge precisa estar aberto no perfil desejado.

  • A sessão do X precisa estar autenticada para buscas completas.

  • O MCP não armazena senha e não contorna login, CAPTCHA ou controles do site.

  • Não é necessário iniciar o Edge com porta de debugging nem usar perfil temporário.

Available Tools

2 tools
list_cua_toolsA

Lista as ferramentas disponiveis no CUA MCP conectado ao Edge.

ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

TDQS

A3.6/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations provided. Description does not disclose that this is a read-only operation or any other behavioral traits beyond listing.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Single concise sentence with no unnecessary words. Front-loaded purpose.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Adequate for a simple listing tool with no output schema. Could mention return format but not essential.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

No parameters exist in schema. Description adds no additional semantics but is sufficient for a parameterless tool.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Description explicitly states 'list the tools' (verb+resource). Clearly distinguishes from sibling 'search_x' which is for searching.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance on when to use this tool versus alternatives. No context provided about prerequisites or intended scenarios.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

search_xA

Pesquisa no X.com usando o Microsoft Edge controlado pelo CUA e retorna o texto acessivel da pagina. Aceita operadores como from:, since:, until:, lang:, filter:links e frases entre aspas.

ParametersJSON Schema
NameRequiredDescriptionDefault
liveNoUsa a aba de posts mais recentes quando possivel
queryYesTermo ou consulta de busca do X

TDQS

A3.9/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description must disclose behaviors. It mentions using Edge controlled by CUA and returning accessible text, but does not describe potential side effects, failure modes, or whether actions like scrolling or logging in occur.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is brief with two sentences—first stating the main purpose, then adding operator details. It is front-loaded and contains no unnecessary words.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Without an output schema, the description should clarify the return format. It only says 'accessible text,' which is vague. Additional details about pagination, scope of results, or limitations would improve completeness.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, but the description adds value by listing accepted search operators (from:, since:, etc.) and quoting, which enhances understanding beyond the schema descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool searches X.com using Microsoft Edge controlled by CUA and returns accessible page text. It is specific about the resource and action, distinguishing it from the sibling tool 'list_cua_tools'.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies usage for searching X.com via CUA-controlled Edge, but provides no explicit guidance on when to use or not use this tool instead of alternatives. No exclusions or context are given.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updatesv0.1.0
    • First observedlist_cua_tools
    • First observedsearch_x
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A3.7/5.0

Scored across 2 tools

Disambiguation5/5

list_cua_tools and search_x have clearly distinct purposes with no overlap. One lists available CUA tools, the other performs searches on X.com, making disambiguation trivial.

Naming Consistency5/5

Both tool names follow a consistent verb_noun pattern in snake_case: 'list_cua_tools' and 'search_x'. The naming convention is uniform across the tool set.

Tool Count3/5

With only 2 tools, the set feels thin for a server named 'X Search MCP' which suggests a broader scope. However, the count is not extreme, and the tools serve basic functions.

Completeness2/5

The server's stated purpose is X search, but it only provides a single search tool and a meta-tool for listing CUA tools. Missing common operations like retrieving individual tweets, user search, or timeline access make the surface incomplete.

Maintenance

ActivityMaintained
ResponsivenessSyncing

Resources

Related MCP Connectors

Related MCP Servers

  • X Search MCP

    konarkm
    A
    license
    B
    quality
    D
    maintenance
    Enables searching X (formerly Twitter) using xAI's Responses API with support for filtering by handles, date ranges, and media understanding, returning structured results with citations.
    1
    11 npm
    1
    MIT

  • Xee-mcp

    Aiyo28
    A
    license
    A
    quality
    D
    maintenance
    Enables AI assistants to search X posts and read user timelines without API costs, using cookie authentication.
    2
    MIT

  • x-mcp-server

    afikrim
    A
    license
    A
    quality
    B
    maintenance
    Enables reading public X (Twitter) content like profiles, tweets, and search results via a stealth browser, without official API costs.
    11
    MIT