Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
作者提供的说明(英文原文)
Model-tiered content pipeline — cheap models seed, expensive models verify
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"sprout": {
"command": "uvx",
"args": [
"sprout-mcp"
]
}
}
}在终端中运行这条命令
claude mcp add sprout -- uvx sprout-mcp粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"sprout": {
"command": "uvx",
"args": [
"sprout-mcp"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"sprout": {
"type": "stdio",
"command": "uvx",
"args": [
"sprout-mcp"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月7日
注册名称已通过 GitHub 账户 mepsopti 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.1.4)于8个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Python (uv)。
同一分类中有:Spec Driven Development, Remote Desktop Commander, Signoff
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Five gates that make an agent's 'done' provable: run them, read the verdict, audit the record.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Drop your Opus bill 80%. Model-tiered content pipeline for MCP — cheap models seed work, expensive models verify it.
Sprout routes tasks to the right model tier automatically. Haiku drafts, Sonnet fact-checks, Opus verifies. Every chunk tracks provenance, confidence, and cost.
uvx sprout-mcp
Or add to Claude Code's MCP config (~/.claude/settings.json):
{
"mcpServers": { "sprout": { "command": "uvx", "args": ["sprout-mcp"] } } }
Haiku (seed) → Sonnet (watered) → Opus (sprouted)
Draft Fact-check Verify $0.005/M $0.015/M $0.075/M
Instead of running everything through Opus at $75/M output tokens, most work stays at Haiku's $5/M. Only the final verification — typically 10-20% of total work — touches Opus.
| Tool | Description |
|---|---|
submit_chunk | Store content with provenance (model, task type, sources) |
get_review_queue | List chunks needing review, filtered by confidence/project |
mark_reviewed | Promote (seed→watered→sprouted) or reject chunks |
recommend_model | Get model recommendation for a task type |
get_stats | Dashboard of chunk counts, confidence levels, token usage |
export_chunks | Export verified chunks as JSON |
opus_test | Generate structured review summary for batch verification |
schedule_task | Schedule tasks to run at a specific time or delay |
list_scheduled | View pending scheduled tasks |
cancel_scheduled | Cancel a pending scheduled task |
configure_routing | Add/update routing rules at runtime |
get_cost_report | Estimated spend per model with real pricing |
retry_on_error | Track failed attempts with backoff guidance |
| Variable | Default | Description |
|---|---|---|
SPROUT_DB_PATH | ~/.sprout/sprout.db | SQLite database location |
SPROUT_CONFIG | *(none)* | Path to JSON config file for custom routes and pricing |
SPROUT_MAX_RETRIES | 3 | Max retry attempts before giving up |
SPROUT_RETRY_BACKOFF | 2.0 | Exponential backoff base (seconds) |
Create a JSON file and point SPROUT_CONFIG at it:
{
"routes": { "code_review": { "tier": "sonnet", "reason": "Code analysis needs reasoning" }, "translation": { "tier": "haiku", "reason": "Straightforward language task" } }, "pricing": { "custom-model": 10.00 } }
| Task Type | Tier | Why |
|---|---|---|
biography_synthesis | haiku | Factual summarization |
council_description | haiku | Historical summarization |
document_synopsis | haiku | Content summarization |
json_validation | haiku | Structural verification |
summarization | haiku | General summarization |
data_extraction | haiku | Structured extraction |
fact_check_first_pass | sonnet | Cross-reference claims |
code_review | sonnet | Code analysis |
fact_check_final | opus | Deep factual verification |
theological_analysis | opus | Domain expertise required |
complex_analysis | opus | Deep reasoning required |
Unknown task types default to haiku — start cheap, escalate if needed.
You: Use recommend_model for "biography_synthesis"
Sprout: biography_synthesis → haiku-4.5 (Factual summarization)
You: Use submit_chunk to store the Haiku output Sprout: Stored chunk abc12345 [seed] for person-001.biography
You: Use get_review_queue to see what needs fact-checking Sprout: 1 chunk pending review
You: Use mark_reviewed to promote after Sonnet fact-checks it Sprout: Chunk abc12345 → watered (verified by sonnet-4.6)
You: Use get_cost_report Sprout: haiku-4.5: ~1,300 tokens (1 chunk) — $0.0065 Total: $0.0065
git clone https://github.com/mepsopti/sprout-mcp.git
cd sprout-mcp uv sync --extra dev uv run pytest
If Sprout saves you money on your AI bill, consider buying me a coffee:
MIT
mcp-name: io.github.mepsopti/sprout-mcp