StageSpirit
Click on "Deploy 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., "@StageSpirit show me the current runtime status and data generation"
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.
StageSpirit 灵幕
StageSpirit 将直播消息、声音、虚拟形象和播出画面连接起来,支持 Agent 参与直播互动,并集中管理服务连接、表达规则和展台配置。
当前提供 D1 首批运行基础,可安装 Python 包、启动 stdio MCP 服务,并通过 runtime_status 查询运行配置、持久数据代际和实例状态。完整功能按实施计划分批交付。
安装与启动
使用 Python 3.12 和 uv 0.8.2,在仓库根目录运行。
uv sync --frozen --no-dev
uv run --frozen --no-dev stagespirit --version
uv run --frozen --no-dev stagespirit serve --profile defaultserve 通过标准输入输出等待 MCP 客户端请求。客户端关闭连接后,服务退出并释放实例锁。日志写入标准错误。安装与开发检查见工程约定。
Related MCP server: Persona MCP Server
MCP 客户端配置
以下配置适用于使用 mcpServers 的 stdio 客户端。将 command 改为实际安装目录下的 .venv\Scripts\stagespirit.exe。macOS 和 Linux 对应 .venv/bin/stagespirit。
{
"mcpServers": {
"stagespirit": {
"command": "C:\\projects\\StageSpirit\\.venv\\Scripts\\stagespirit.exe",
"args": ["serve", "--profile", "default"]
}
}
}连接后发现 runtime_status,以 {} 调用即可读取状态。data_generation 随当前运行配置持久保存,instance_id 在每次启动时更新。
使用 --data-dir 可选择应用数据根目录。默认使用系统用户数据目录,Windows 为 %LOCALAPPDATA%\StageSpirit。每个 profile 的状态位于 profiles/<profile_id>/,同一 profile 同时运行一个实例。
项目结构
目录 | 用途 |
src/stagespirit_mcp/ | Python 后端源码 |
web/stage/ | 展台前端的规划位置 |
docs/ | 产品、需求、架构、设计、参考、指南和测试文档 |
contracts/ | MCP 工具定义和共享 Schema |
examples/contracts/ | 接口请求与响应示例 |
tests/ | 测试及测试数据 |
scripts/ | 开发维护脚本 |
参与项目
通过仓库 Issues 报告问题、提出功能需求和讨论设计,按贡献指南提交变更。安全问题使用私密报告入口。
项目仓库 · 问题与设计讨论 · 贡献指南 · 安全报告方式
许可
本仓库自有源码、文档、Schema 和示例采用 Apache License 2.0。第三方材料在各自位置保留来源、版权和适用许可。
Available Tools
1 toolruntime_status运行状态ARead-onlyIdempotent
读取当前运行配置、数据代际、实例身份和已启动模块的固定摘要。
Inputs: no named parameters. Returns: state, modules, version, uptime_ms, +6 more
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 1 tool
With a single tool there is no possibility of misselection or overlap. The purpose (reading a fixed runtime summary) is unambiguous.
runtime_status follows a clear snake_case noun_topic pattern that is readable and predictable. There is no second name to compare against, so consistency is trivially satisfied but the convention itself is sound.
A single tool is too thin for a server that appears to expose runtime/configuration state; agents will hit a dead end immediately. The surface needs at least a few more inspection operations to justify a server.
The tool bundles config, data generation, instance identity, and module summary into one read-only call with no way to query individual modules, refresh, or trigger any lifecycle action. Only one direction (read) of the domain is covered.
Maintenance
Related MCP Connectors
Agent-first social media publishing, scheduling, analytics, and unified inbox API + 16-tool MCP server across 18 social platforms (X, LinkedIn, Instagram, YouTube, TikTok, Threads, Bluesky, Reddit, Telegram, Discord, Mastodon, Facebook, Pinterest).
Build and manage AI-native customer support agents from Claude or any MCP client.
Build and run AI workflows, agents, data tables and 1000+ API integrations from any MCP client.
Agent communication platform for agent to agent messaging via MCP. Messages, channels, skills.
Related MCP Servers
- AlicenseAqualityBmaintenanceAI-driven AR lens orchestrator for live OBS streams that enables control of Streamfog face filters, AR effects, and Vtuber avatars through MCP tools via the local Streamer.bot WebSocket bridge.52MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to control a desktop virtual character (VRM) by playing animations, showing/hiding the character, and checking runtime status through the MCP protocol.768,733 npm2MIT
- AlicenseNot gradedqualityBmaintenanceEnables voice-first interactions with AI agents and MCP tools, supporting speech input/output, STT/TTS, and a provider-independent agent core.1MIT
- AlicenseNot gradedqualityBmaintenanceEnables local, user-owned AI Avatar setup and orchestration, connecting AI hosts via stdio MCP with secure OAuth and privacy controls.MIT