OpenAI MCP Server
# OpenAI MCP Server
A Model Context Protocol server that provides tools to manage OpenAI API keys & spending.
> [!CAUTION]
> This server can access the OpenAI API and may represent a security risk. Exercise caution when using this MCP server to ensure this does not expose any sensitive data.
## Available Tools
TODO: Add tool info here
## Installation
### Using uv (recommended)
When using [`uv`](https://docs.astral.sh/uv/) no specific installation is needed. We will use [`uvx`](https://docs.astral.sh/uv/guides/tools/) to directly run *mcp-server-openai*.
### Using PIP
Alternatively you can install `mcp-server-openai` via pip:
```
pip install mcp-server-openai
```
After installation, you can run it as a script using:
```
python -m mcp_server_openai
```
## Configuration
### Configure for Claude.app
Add to your Claude settings:
<details>
<summary>Using uvx</summary>
```json
{
"mcpServers": {
"openai": {
"command": "uvx",
"args": ["mcp-server-openai"],
"env": {
"OPENAI_ADMIN_API_KEY": "your_openai_admin_api_key"
}
}
}
}
```
</details>
<details>
<summary>Using docker</summary>
```json
{
"mcpServers": {
"openai": {
"command": "docker",
"args": ["run", "-i", "--rm", "mcp/openai"],
"env": {
"OPENAI_ADMIN_API_KEY": "your_openai_admin_api_key"
}
}
}
}
```
</details>
<details>
<summary>Using pip installation</summary>
```json
{
"mcpServers": {
"openai": {
"command": "python",
"args": ["-m", "mcp_server_openai"],
"env": {
"OPENAI_ADMIN_API_KEY": "your_openai_admin_api_key"
}
}
}
}
```
</details>
## Debugging
You can use the MCP inspector to debug the server. For uvx installations:
```
npx @modelcontextprotocol/inspector uvx mcp-server-openai
```
Or if you've installed the package in a specific directory or are developing on it:
```
cd path/to/servers/src/openai
npx @modelcontextprotocol/inspector uv run mcp-server-openai
```
## Contributing
We encourage contributions to help expand and improve mcp-server-openai. Whether you want to add new tools, enhance existing functionality, or improve documentation, your input is valuable.
Pull requests are welcome! Feel free to contribute new ideas, bug fixes, or enhancements to make mcp-server-openai even more powerful and useful.
## License
mcp-server-openai is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: get_costs retrieves financial data for a specified period, while get_projects fetches organizational project information. There is no overlap in functionality, making it easy for an agent to select the correct tool based on the task.
Both tools follow a consistent verb_noun pattern (get_costs and get_projects), using the same verb 'get' and snake_case formatting. This uniformity makes the tool set predictable and easy to understand.
With only two tools, the server feels thin and under-scoped for an OpenAI MCP server, which might be expected to handle more operations like managing models, generating completions, or listing usage. The limited count suggests incomplete coverage of the domain.
The tool set is severely incomplete for an OpenAI server, lacking core operations such as creating or managing projects, generating text or images, or accessing model details. This will likely cause agent failures when attempting common OpenAI-related tasks.