Leverages OpenAI's GPT-5.1 model with built-in Web Search capabilities to perform research queries and return results with citations through the Responses API.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@GPT Research MCP Server research the latest developments in quantum computing for 2025"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
OpenAI GPT-5.1 の Responses API と組み込み Web Search ツールを活用したリサーチ MCP サーバーです。
research(query): GPT-5.1 の Web Search 機能を使用してクエリを調査し、引用情報付きの結果を返します
Related MCP server: MCP Deep Search
Python 3.12+
uv パッケージマネージャー
OpenAI API キー
git clone https://github.com/your-username/gpt-research-mcp.git
cd gpt-research-mcpuv syncexport OPENAI_API_KEY="sk-..."OpenAI 互換 API（Azure OpenAI、ローカル LLM など）を使用する場合は、OPENAI_BASE_URL を設定します:
export OPENAI_BASE_URL="https://your-custom-endpoint.com/v1"設定しない場合は、デフォルトの https://api.openai.com/v1/ が使用されます。
LangFuse を使用して OpenAI API 呼び出しの可観測性を有効にできます。以下の環境変数をすべて設定すると自動的に有効化されます:
export LANGFUSE_SECRET_KEY="sk-lf-..."
export LANGFUSE_PUBLIC_KEY="pk-lf-..."
export LANGFUSE_BASE_URL="https://cloud.langfuse.com" # EU region
# export LANGFUSE_BASE_URL="https://us.cloud.langfuse.com" # US region環境変数が設定されていない場合は、通常の OpenAI SDK が使用されます。
claude mcp add gpt-research -- uv run --directory /path/to/gpt-research-mcp main.pyまたは ~/.claude/settings.json に直接追加:
{
"mcpServers": {
"gpt-research": {
"command": "uv",
"args": ["run", "--directory", "/path/to/gpt-research-mcp", "main.py"],
"env": {
"OPENAI_API_KEY": "sk-...",
"OPENAI_BASE_URL": "https://api.openai.com/v1/"
}
}
}
}OPENAI_BASE_URL はオプションです。カスタムエンドポイントを使用しない場合は省略可能です。
/path/to/gpt-research-mcp を実際のパスに置き換えてください。
uv run python main.pyMCP クライアント（Claude Desktop 等）から research ツールを呼び出します:
research("2024年の AI トレンドについて教えてください")research(query: str) -> strパラメータ | 型 | 説明 |
|
| 調査したい質問やトピック |
戻り値: GPT-5.1 が生成したリサーチ結果（引用情報を含む）
コンポーネント | 技術 |
パッケージ管理 | uv |
Python バージョン | 3.12 |
MCP フレームワーク | FastMCP 2.0 |
OpenAI クライアント | openai (公式ライブラリ) |
モデル | gpt-5.1 |
トレーシング（オプション） | LangFuse |
Ask questions, search for information, or consult about complex problems in English.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It mentions the tool can 'ask questions, search for information, or consult,' but fails to describe key traits like whether it performs read-only operations, requires authentication, has rate limits, or what the output entails. This leaves significant gaps in understanding how the tool behaves beyond its basic purpose.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise and front-loaded in a single sentence, with no wasted words. It efficiently states the tool's purpose and language requirement. However, it could be more structured by separating key points, but its brevity is appropriate given the simple tool nature.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity is low (1 parameter, no siblings) and an output schema exists (which should cover return values), the description is somewhat complete but has gaps. It covers the basic purpose and language, but lacks behavioral details and parameter guidance. With no annotations, it should do more to compensate, making it minimally adequate but not fully helpful.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 1 parameter with 0% description coverage, meaning the schema provides no details about the 'query' parameter. The description adds minimal semantics by implying the query should be in English and related to questions or complex problems, but it doesn't specify format, constraints, or examples. This is insufficient to compensate for the low schema coverage, resulting in poor parameter understanding.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states the tool's purpose as 'Ask questions, search for information, or consult about complex problems in English,' which is clear but vague. It specifies the action (ask/search/consult) and language (English) but lacks specificity about what resources or domains it operates on, and with no siblings, differentiation is not needed. However, 'research' as a name is broad, and the description doesn't fully clarify its scope beyond general information-seeking.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no explicit guidance on when to use this tool versus alternatives. It mentions 'complex problems' and 'English,' which imply some context, but there are no stated exclusions, prerequisites, or comparisons to other tools. With no sibling tools, the lack of guidance is less critical, but it still offers minimal direction for effective usage.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
With only one tool, there is no possibility of ambiguity or overlap between tools. The single tool 'research' has a clearly defined purpose, so an agent cannot misselect between non-existent alternatives.
A single tool inherently exhibits perfect naming consistency, as there are no other tools to compare it against for patterns or conventions. The name 'research' is straightforward and follows a simple verb-based naming style.
One tool is too few for a server named 'GPT Research MCP Server', which suggests a research-oriented domain that typically requires multiple operations (e.g., search, summarize, cite, query). A single tool feels thin and under-scoped for such a purpose, limiting functionality.
The tool surface is severely incomplete for a research server. While 'research' covers general inquiry, there are obvious gaps such as specialized search, result filtering, citation management, or data analysis tools that would be expected in this domain, likely causing agent failures in complex tasks.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Web search, fetch, extract, and research for AI agents. Markdown output + AI-synthesized answers.
Live AI-native web search with citations. One tool for every MCP client. Flat per-request pricing.
Web research for agents: quality-scored Google search, webpage extraction, and deep research.
8-tool AI web intelligence suite: search, scrape, screenshot, SEO, docs, crypto, code.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/testusuke/gpt-research-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
