Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
TOP_PNo核采样阈值0.95
LUMA_DEBUGNo`1`/`true` 时等同开启 `INCLUDE_META`
MAX_TOKENSNo最大生成 token 数(部分模型有硬上限)8192
MODEL_NAMENo模型名称覆盖(默认自动选择)
MULTI_CROPNo大图多裁剪,设为 `false` 关闭true
CUSTOM_PATHNo可选:API 路径/chat/completions
TEMPERATURENo采样温度0.7
INCLUDE_METANo为 `true` 时在工具结果末尾附加预处理/API 耗时等 metafalse
MCP_HTTP_HOSTNoHTTP 模式监听地址(Docker 内需为 `0.0.0.0`)0.0.0.0
MCP_HTTP_PORTNoHTTP 模式监听端口3000
MCP_TRANSPORTNo传输方式:`stdio`(默认)或 `http`(Streamable HTTP)stdio
ZHIPU_API_KEYNoZhipu 提供商的 API Key(默认模型 glm-4.6v)
CUSTOM_API_KEYNoCustom 提供商的 API Key,用于任意 OpenAI 兼容端点
MCP_HTTP_TOKENNoHTTP 模式 Bearer token;局域网共享务必设置(默认为空,不鉴权)
MODEL_PROVIDERNo模型提供商:`zhipu`、`siliconflow`、`qwen`、`volcengine`、`hunyuan`、`custom`zhipu
CUSTOM_BASE_URLNoCustom 提供商的 OpenAI 兼容 API 地址(例如 https://your-endpoint.com/v1)
ENABLE_THINKINGNo思考模式,设为 `false` 关闭true
HUNYUAN_API_KEYNoHunyuan 提供商的 API Key(默认模型 hunyuan-t1-vision-20250916)
CUSTOM_MODEL_NAMENoCustom 提供商的模型名称
CUSTOM_TIMEOUT_MSNo可选:超时毫秒60000
DASHSCOPE_API_KEYNoQwen 提供商的 API Key(默认模型 qwen3-vl-flash)
BASE_VISION_PROMPTNo自定义基础视觉提示词(设为空字符串可关闭,默认使用内置默认值)
CUSTOM_AUTH_HEADERNo可选:`bearer` / `x-api-key` / `custom`(默认 bearer)
VOLCENGINE_API_KEYNoVolcengine 提供商的 API Key(默认模型 doubao-seed-1-6-flash-250828)
SILICONFLOW_API_KEYNoSiliconFlow 提供商的 API Key(默认模型 deepseek-ai/DeepSeek-OCR)
CUSTOM_THINKING_MODENo可选:`disabled` / `openai` / `qwen_extra_body`disabled
MULTI_CROP_MAX_TILESNo多裁剪最大图块数(含原图,1–16)5
CUSTOM_AUTH_HEADER_VALUENo可选:自定义 Header 模板,例如 "X-API-Key: {{key}}"(`{{key}}` 会被替换为 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

NameDescription

No tools

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources