Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
AUTODL_TOKENYesAutoDL control panel login token (raw token, do not add Bearer prefix).
AUTODL_CONFIGNoPath to an alternative config.yaml file.

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

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}

Tools

Functions exposed to the LLM to take actions

NameDescription
power_onA

为指定 UUID 开机(会计费);缺货后台持续重试,立即返回任务 ID。

power_on_no_gpuA

无卡模式开机(仍会计费,无 GPU);仅支持按量计费实例,立即返回后台任务 ID。不会自动将已运行的 GPU 实例切换为无卡模式,请先关机并确认状态。

power_offA

关机(中断实例上的运行任务);先取消开机重试,返回接受结果而非最终状态。

instance_statusA

查询指定 UUID 的实例状态与配置规格:GPU 型号、单卡显存 GB(1024³ 字节)、分配的 CPU 核心数;无卡模式不代表 GPU 已挂载。不返回密码或 SSH 凭证。

list_instancesA

批量查询 1–100 个实例 UUID 的状态和硬件规格,共享一次列表分页。返回 instances 与逐项 errors;价格来源未验证时 price 为 null,不代表免费。不会返回密码或 SSH 凭证。

switch_modeA

切换 GPU/无卡模式,立即返回后台任务 ID:关机、确认关机、按目标模式重新开机。会中断运行任务,重新开机会计费;执行前必须确认。无卡目标仅支持按量计费。不是可回滚事务;失败或取消可能停在关机状态,不自动恢复。任务不跨服务重启恢复。

job_statusB

查询当前进程中的开机/切换任务、尝试次数、累计耗时、重试时间和错误。next_retry_in_seconds 为快照时距下次重试的剩余秒数(向上取整,无重试时为 null)。

wait_for_jobA

只读等待当前进程任务结束、缺货或超时后返回快照;默认缺货立即返回。timeout_seconds 为 0–300 秒(默认 15,0 立即返回)。return_on_capacity=False 可等待穿过缺货阶段;活动 phase 表示尚未结束。缺货返回后先汇报 attempts/elapsed_seconds/next_retry_in_seconds,不要紧密重复查询。next_retry_in_seconds 为快照时距下次重试的剩余秒数(向上取整,无重试时为 null)。本次等待超时或断开不会取消后台开机任务;任务不跨服务重启恢复。

cancel_jobA

取消开机/切换任务的后续步骤,不撤销已接受的电源请求,不自动关机或恢复。

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A3.8/5.0

Scored across 9 tools

Disambiguation4/5

Tools target mostly distinct actions, but power_on, power_on_no_gpu, and switch_mode all involve powering on or cycling instances and could be misselected without reading details. The descriptions clarify scope (e.g., no-card vs GPU, mode switching).

Naming Consistency4/5

Names are mostly snake_case with recognizable verb/noun components, but conventions vary (power_on, job_status, wait_for_job, instance_status). Still predictable and readable overall.

Tool Count5/5

9 tools is well within the ideal range for a focused power and status management server. Each tool covers a specific operation without obvious redundancy.

Completeness4/5

Covers power on/off, no-GPU mode, mode switching, task control, and single/batch status. Broader lifecycle operations (create/delete instance) are missing, but those may be outside the server's apparent scope.

Maintenance

ActivityMaintained
ResponsivenessNo issues