Distillery

io.github.norrietaylor由 GitHub 账号 norrietaylor 发布v0.7.0更新于 Oct 24, 2026

Knowledge base for Claude Code with semantic search, feed intelligence, and teams

已验证STDIO仅桌面效率与工作流知识与记忆开发者工具

概览

AI 生成的概览

一个本地团队知识库，让助手能够记录、分类并语义检索笔记、会议纪要和书签。

可以做什么
Distillery 将提炼后的团队知识以向量嵌入形式存储在本地 DuckDB 数据库中，并通过自然语言检索。它提供 Claude Code 斜杠命令，用于记录会话知识、带来源的语义召回、多条目综合、会议纪要、带摘要的书签、信息源监控与摘要、GitHub issue 与 PR 同步，以及团队知识仪表盘。它也可以作为带 GitHub OAuth 的托管 HTTP 服务运行，供团队访问。
适用场景
当团队希望拥有一个私密、可检索的记忆库，用来保存决策、会议和链接，并让助手查询和综合时，值得添加。适合反复召回过往上下文、监控信息源和汇总 GitHub 活动。一次性提问或公开网络查询则不需要。
运行要求
作为 Python 包在本地通过 stdio 运行（需要 Python 3.11+ 和 uv）；默认的 fastembed 提供方离线运行，无需 API 密钥。可选的托管嵌入提供方需要 DISTILLERY_EMBEDDING_PROVIDER 以及 JINA_API_KEY 或 OPENAI_API_KEY。提供用于评估的托管演示端点。仅支持桌面端。
安装前请注意
托管演示端点仅供评估，不应存放敏感或机密数据。可选的嵌入提供方会把内容发送给 Jina 或 OpenAI，并需要 JINA_API_KEY 或 OPENAI_API_KEY 密钥。该服务器会把知识条目、书签、纪要和同步的 GitHub 数据写入本地存储，信息源或 GitHub 同步会从第三方来源拉取数据。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Distillery，将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Desktop only，通过 STDIO。 STDIO 服务会启动本地进程，因此需要 SourceWeft 桌面宿主。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

[Distillery]

Distillery

Team Knowledge, Distilled
Capture, classify, connect, and surface team knowledge through conversational commands.

Documentation · Skills · Quick Start · Roadmap · Slides

[PyPI version] [PyPI downloads] [License] [Python version]

What is Distillery?

Distillery is a team knowledge base accessed through Claude Code skills. It refines raw information from working sessions, meetings, bookmarks, and conversations into concentrated, searchable knowledge — stored as vector embeddings in DuckDB and retrieved through natural language. Runs locally over stdio or as a hosted HTTP service with GitHub OAuth for team access.

Distillery captures the highest-value transformation — from noise to signal — and makes it a tool the whole team can use.

Full documentation: norrietaylor.github.io/distillery

[Distillery demo — /distill captures a decision, /pour synthesizes it]

Skills

Distillery provides 15 Claude Code slash commands:

SkillPurposeExample
/distillCapture session knowledge with dedup detection/distill "We decided to use DuckDB for local storage"
/recallSemantic search with provenance/recall distributed caching strategies
/pourMulti-entry synthesis with citations/pour how does our auth system work?
/compassContrast internal knowledge vs. ambient intelligence for a directional assessment/compass our caching strategy
/bookmarkStore URLs with auto-generated summaries/bookmark https://example.com/article #caching
/minutesMeeting notes with append updates/minutes --update standup-2026-03-22
/classifyClassify entries and triage review queue/classify --inbox
/watchManage monitored feed sources/watch add github:duckdb/duckdb
/radarAmbient feed digest with source suggestions/radar --days 7
/tuneAdjust feed relevance thresholds/tune relevance 0.4
/digestTeam activity summary from internal entries/digest --days 7 --project myapp
/gh-syncSync GitHub issues/PRs into the knowledge base/gh-sync owner/repo --issues
/investigateDeep context builder with relationship traversal/investigate distributed caching
/briefingTeam knowledge dashboard with metrics/briefing --days 7
/setupOnboarding wizard for MCP connectivity and config/setup

Quick Start

Step 1: Install the Plugin

bash
claude plugin marketplace add norrietaylor/distilleryclaude plugin install distillery

This installs all 15 skills. The plugin does not configure an MCP server automatically — run /setup (below) to add one. The recommended setup runs locally via uvx --from 'distillery-mcp[fastembed]>=0.6.0' distillery-mcp — a private, self-contained knowledge base on your machine, with on-device fastembed embeddings (no API key required). Requires Python 3.11+ and uv (install: curl -LsSf https://astral.sh/uv/install.sh | sh).

Step 2 (Optional): Use Jina or OpenAI Instead

The default fastembed provider runs offline with no API key. If you'd rather use a hosted embedding service, set DISTILLERY_EMBEDDING_PROVIDER and provide the matching API key:

bash
# Jina (free tier at jina.ai)export DISTILLERY_EMBEDDING_PROVIDER=jinaexport JINA_API_KEY=jina_...
# Or OpenAIexport DISTILLERY_EMBEDDING_PROVIDER=openaiexport OPENAI_API_KEY=sk-...

uvx inherits these from your shell environment. See distillery.yaml.example for the full provider configuration block (including Option C for fastembed model selection).

Restart Claude Code and run the onboarding wizard:

text
/setup

Try the Hosted Demo (Opt-In)

Want to evaluate without installing anything locally? Configure the hosted demo at distillery-mcp.fly.dev instead of a local server:

bash
claude mcp add distillery --scope user --transport http --url https://distillery-mcp.fly.dev/mcp

Demo Server: distillery-mcp.fly.dev is for evaluation only. Do not store sensitive or confidential data.

See the Local Setup Guide for full configuration options, or deploy your own instance for team use.

Development

bash
uv pip install -e ".[dev]"# orpip install -e ".[dev]"pytest                              # run testsmypy --strict src/distillery/       # type checkruff check src/ tests/              # lint

See Contributing for the full guide.

License

Apache 2.0 — see LICENSE for details.

来源：README.md，提交 d6697cc

工具

0
SourceWeft 尚未在沙箱中启动过该软件包，因此其工具未知。

版本历史

1
  1. v0.7.0最新Sep 16, 2026

评分

已满足适用检查项的 40%

该评分衡量此收录条目的完整程度，依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价，也不表示服务器安全或当前可用：可连通性是带日期的检查结果，评分也不包含安装量或用户评价。

  • 近 30 天内可连通25 分

    请保持端点能够响应 MCP 握手；SourceWeft 每 30 天探测一次。

  • 已声明认证方式10 分

    请在注册中心条目中声明凭据请求头，或保持端点能够响应探测，以便判定其认证方式。

  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list，或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分

    请在仓库或注册中心条目中声明许可证。