Luma MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TOP_P | No | Nucleus sampling threshold. Default is 0.95. | |
| LUMA_DEBUG | No | Set to 1 or true to enable INCLUDE_META behavior. Default is off. | |
| MAX_TOKENS | No | Maximum number of generated tokens. Default is 8192. SiliconFlow DeepSeek-OCR is internally capped at 4096. | |
| MODEL_NAME | No | Override the model name. Defaults to an automatically selected model per provider. | |
| MULTI_CROP | No | Enable multi-crop for large images; set to false to disable. Default is true. | |
| CUSTOM_PATH | No | Custom endpoint API path. Default is /chat/completions. | |
| TEMPERATURE | No | Sampling temperature. Default is 0.7. | |
| INCLUDE_META | No | Set to true to append preprocessing and API timing metadata to the tool result. Default is false. | |
| MCP_HTTP_HOST | No | HTTP mode listen host. Default is 0.0.0.0. | |
| MCP_HTTP_PORT | No | HTTP mode listen port. Default is 3000. | |
| MCP_TRANSPORT | No | Transport mode: stdio (default) or http for Streamable HTTP. | |
| ZHIPU_API_KEY | No | Zhipu API key. Required when using the zhipu provider, which is the default. | |
| CUSTOM_API_KEY | No | API key for a custom OpenAI-compatible endpoint. Required when MODEL_PROVIDER is custom. | |
| MCP_HTTP_TOKEN | No | Bearer token for HTTP mode. Strongly recommended when sharing over a LAN. Default is empty, meaning no authentication. | |
| MODEL_PROVIDER | No | Model provider: zhipu, siliconflow, qwen, volcengine, hunyuan, or custom. Defaults to zhipu. | |
| CUSTOM_BASE_URL | No | Base URL for a custom OpenAI-compatible endpoint. Required when MODEL_PROVIDER is custom. | |
| ENABLE_THINKING | No | Thinking mode; set to false to disable. Default is true. | |
| HUNYUAN_API_KEY | No | Hunyuan API key. Required when MODEL_PROVIDER is hunyuan. | |
| CUSTOM_MODEL_NAME | No | Model name for a custom OpenAI-compatible endpoint. Required when MODEL_PROVIDER is custom. | |
| CUSTOM_TIMEOUT_MS | No | Custom endpoint request timeout in milliseconds. Default is 60000. | |
| DASHSCOPE_API_KEY | No | Qwen/DashScope API key. Required when MODEL_PROVIDER is qwen. | |
| BASE_VISION_PROMPT | No | Custom base vision prompt. Set to an empty string to disable. | |
| CUSTOM_AUTH_HEADER | No | Custom endpoint auth header style: bearer, x-api-key, or custom. Default is bearer. | |
| VOLCENGINE_API_KEY | No | Volcengine API key. Required when MODEL_PROVIDER is volcengine. | |
| SILICONFLOW_API_KEY | No | SiliconFlow API key. Required when MODEL_PROVIDER is siliconflow. | |
| CUSTOM_THINKING_MODE | No | Custom endpoint thinking mode: disabled, openai, or qwen_extra_body. Default is disabled. | |
| MULTI_CROP_MAX_TILES | No | Maximum number of crop tiles including the original image, from 1 to 16. Default is 5. | |
| CUSTOM_AUTH_HEADER_VALUE | No | Custom header template, for example 'X-API-Key: {{key}}'. The {{key}} placeholder is replaced with the API key. |
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
Server capabilities have not been inspected yet.
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
No tools | |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
This server cannot be deployed
Maintenance
ActivityActive
ResponsivenessNo issues