Journal Mandala MCP
Journal Mandala MCP
这是一个用于从 Claude 直接向 Journal Mandala 写入日记条目的 MCP 服务器。
当你希望在对话过程中将“思维流”记录到 Mandala 时使用。
必要条件
Node.js 20+
Claude Desktop (或 Claude Code)
Journal Mandala 的 API 密钥(获取方法请参考“获取 API 密钥”章节)
Related MCP server: Obsidian Nexus
安装
使用 npm 进行全局安装。
npm install -g --install-links github:takezone/journal-mandala-mcp--install-links 是必须的(如果不加,会变成指向临时目录的软链接,导致无法启动)。
更新时也执行相同的命令。如果过去在没有 --install-links 的情况下安装导致损坏,请执行:
npm uninstall -g journal-mandala-mcp
# 残骸掃除 (npm prefix は `npm prefix -g` で確認)
rm -rf $(npm prefix -g)/lib/node_modules/journal-mandala-mcp
rm -rf $(npm prefix -g)/lib/node_modules/.journal-mandala-mcp-*
npm install -g --install-links github:takezone/journal-mandala-mcpClaude Desktop 配置
编辑 ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
单账号运行
{
"mcpServers": {
"journal-mandala": {
"command": "journal-mandala-mcp",
"env": {
"JOURNAL_MANDALA_API_KEY": "jm_XXXXXXXXXXXXXXXXXXXXXXXX"
}
}
}
}需要切换 2 个账号(个人 / 工作)的情况
{
"mcpServers": {
"journal-mandala-personal": {
"command": "journal-mandala-mcp",
"env": {
"JOURNAL_MANDALA_API_KEY": "jm_PERSONAL_KEY"
}
},
"journal-mandala-work": {
"command": "journal-mandala-mcp",
"env": {
"JOURNAL_MANDALA_API_KEY": "jm_WORK_KEY"
}
}
}
}在 Claude 中它们会显示为不同的工具。可以通过自然语言切换,例如“记录到工作”。
配置完成后请重启 Claude Desktop。
获取 API 密钥
目前请将以下信息告知 Journal Mandala 管理员(@takezone)以进行签发:
你的 Google 账号邮箱地址
标签名称(例如:
alice-personal)
收到管理员发送的 API 密钥后,将其复制粘贴到上述配置的 JOURNAL_MANDALA_API_KEY 中。
注意: 未来计划支持在浏览器中自动签发。
使用方法
在与 Claude 的对话中,请求如“将今天发生的事情记录到日记中”,add_journal_entry 工具就会被调用,并将条目添加到 Journal Mandala。
提供的工具:
add_journal_entry — 添加日记条目
title标题(可选)event事件(发生了什么)thought思考(感受如何)entry_date日期YYYY-MM-DD(省略时默认为 JST 今天)entry_time时间HH:MMtags标签数组 — 必须作为字符串数组传递✅ 正确:
["工作", "运动"]❌ 错误:
["[\"工作\",\"运动\"]"](不要将 JSON 字符串化的数组作为单个元素)
star_rating1-5 评分
find_entries — 最近条目列表
用于获取 ID。按最近更新顺序返回。
limit获取数量(默认 20,最大 100)
update_entry — 更新现有条目
通过指定 ID 进行部分更新。仅覆盖指定的字段,省略的字段保持不变。
id必须其他字段与
add_journal_entry/add_todo的所有字段均可任意指定
使用方法:通过 find_entries 查询 ID → 使用 update_entry 更新
add_todo — 添加待办事项
title任务名称(必须)thought关于任务的备注(可选)entry_date创建日期(省略时默认为 JST 今天)due_date截止日期YYYY-MM-DD(可选)todo_statuspending|in_progress|completed|cancelled(省略时默认为 pending)todo_points预估点数(可选)is_today今日待办列表标记(可选)tags标签数组1: 微妙/负面的事件2: 普通的一天(默认)3: 发生了还不错的事情的一天4: 发生了相当不错的事情的一天5: 人生重大事件级别不要轻易使用 5。基本使用 2,有好事发生时使用 3 或 4
title / event / thought 中至少需要包含一项。
可选环境变量
JOURNAL_MANDALA_API_URL— API 端点 (默认:https://journal-mandala.vercel.app)
故障排除
Claude 中不显示工具 → 可能是配置文件 JSON 语法错误。请在 Claude Desktop 的 Developer 菜单中查看 MCP 日志
API 错误 401 →
JOURNAL_MANDALA_API_KEY无效。请请求管理员重新签发网络错误 → 请检查 Vercel 的运行状态
管理员指南:API 密钥的管理与恢复
供运营者在丢失密钥时参考。由于此 MCP 引用的是保存在 Vercel 侧的正本,即使 ~/.claude.json 的副本丢失也可以恢复。
3 层结构
层级 | 位置 | 角色 |
服务器 env var | Vercel | 初始 seed 的静态账号数组(正本) |
服务器 KV | Vercel KV (Upstash Redis) | 通过 |
客户端 |
| 用于认证的密钥副本(丢失可重新获取) |
API 认证时按 KV → env var 的顺序进行核对。
恢复步骤(项目所有者)
在项目目录(journal-mandala 本体)中:
# Vercel プロジェクトと連携(初回のみ)
vercel link --yes --project journal-mandala
# サーバー環境変数をローカルに取得
vercel env pull .env.vercel
# env var 内のキー確認
grep JM_API_ACCOUNTS .env.vercel
# KV 内のキー一覧スキャン
source .env.vercel
curl -s "${KV_REST_API_URL}/scan/0/match/jm:account:*/count/100" \
-H "Authorization: Bearer ${KV_REST_API_TOKEN}"
# 特定アカウントの中身を見る
curl -s "${KV_REST_API_URL}/get/jm:account:jm_XXXXX" \
-H "Authorization: Bearer ${KV_REST_API_TOKEN}"
# 終わったら機密情報を必ず削除
rm -f .env.vercel .vercel/project.json && rmdir .vercel 2>/dev/null签发新密钥
在 Chrome 中登录目标 Google 账号
打开
https://journal-mandala.vercel.app/api/enroll?label=<任意标签>OAuth 授权 → 屏幕上会显示一次 API 密钥,请复制
反映到
~/.claude.json的mcpServers.*.env.JOURNAL_MANDALA_API_KEY中
使用同一个 Google 账号重新 enroll 时,旧密钥会自动失效并替换为新密钥。
在其他 Mac(第 2 台及以后)上进行设置
由于 API 密钥与 Google 账号绑定,即使更换机器也可以使用相同的密钥(无需重新 enroll)。
# 1. (未インストールなら) Node.js をインストール
brew install node
# 2. MCPサーバーをグローバルインストール
npm install -g --install-links github:takezone/journal-mandala-mcp
# 3. ~/.claude.json に mcpServers を追記(既存設定を壊さないよう python でマージ推奨)
python3 - <<'PY'
import json, os
p = os.path.expanduser('~/.claude.json')
with open(p) as f: c = json.load(f)
c.setdefault('mcpServers', {})
c['mcpServers'].update({
'journal-mandala-personal': {
'command': 'journal-mandala-mcp',
'env': {'JOURNAL_MANDALA_API_KEY': 'jm_XXXXX'} # ←実キー
},
'journal-mandala-work': {
'command': 'journal-mandala-mcp',
'env': {'JOURNAL_MANDALA_API_KEY': 'jm_YYYYY'} # ←実キー
},
})
with open(p, 'w') as f: json.dump(c, f, indent=2, ensure_ascii=False)
print('updated')
PY
# 4. Claude Code を再起動(/exit → claude 再実行)如果手头没有现有密钥: 请从家里 Mac 的 ~/.claude.json 中复制,或者通过上述“恢复步骤”从 Vercel 获取。
推荐分布式运营: 如果在家里 Mac 和公司 Mac 上使用不同的密钥(例如:?label=personal-home / ?label=personal-work),即使其中一个泄露也不会影响另一个,且轮换时只需分别处理。虽然使用同一个密钥更方便,但可以根据喜好选择。
许可证
MIT
Available Tools
1 tooladd_journal_entryAdd Journal Mandala entryB
Journal Mandala に新しい日記エントリを作成する。event (出来事: 何があったか) と thought (思考: それについてどう感じたか) は両方とも任意だが、少なくとも title/event/thought のいずれかは必須。日付を省略すると日本時間での今日の日付になる。
| Name | Required | Description | Default |
|---|---|---|---|
| title | No | エントリのタイトル(任意。省略時は先頭20文字から自動生成) | |
| event | No | 出来事: 何があったかを事実ベースで記述 | |
| thought | No | 思考: それについてどう感じたか・何を考えたか | |
| entry_date | No | YYYY-MM-DD 形式。省略時は日本時間での今日 | |
| entry_time | No | HH:MM 形式(任意) | |
| tags | No | タグの配列(任意) | |
| star_rating | No | 1-5 の評価(任意) |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It describes key behaviors: the tool creates a new entry (implying a write/mutation operation), specifies default values for title (auto-generated from first 20 characters) and date (today's Japan time), and mentions the requirement for at least one of title/event/thought. However, it lacks details on permissions, error handling, or response format, leaving gaps for a mutation tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is appropriately concise with two sentences that efficiently convey the tool's purpose and key constraints. It is front-loaded with the main action and avoids unnecessary details, though it could be slightly more structured by separating usage rules into bullet points for clarity.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (7 parameters, mutation operation) and lack of annotations and output schema, the description is moderately complete. It covers the creation action, parameter requirements, and defaults, but misses details on permissions, error cases, and what the tool returns, which are important for a write tool with no structured output information.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema description coverage is 100%, providing detailed descriptions for all 7 parameters. The description adds minimal value beyond the schema by mentioning that event and thought are optional but at least one of title/event/thought is required, and that the date defaults to today's Japan time if omitted. This aligns with the baseline of 3 when the schema does most of the work.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Journal Mandala に新しい日記エントリを作成する' (creates a new journal entry in Journal Mandala). It specifies the verb (create/作成する) and resource (journal entry/日記エントリ). However, since there are no sibling tools mentioned, it cannot demonstrate differentiation from alternatives, preventing a score of 5.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides implied usage guidance by stating that at least one of title/event/thought is required, and that the date defaults to today's date in Japan time if omitted. However, it lacks explicit guidance on when to use this tool versus alternatives (none mentioned) or any prerequisites or exclusions, keeping it at a basic level.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
1 tool update
v0.1.0- First observed
add_journal_entry
TDQS
Scored across 1 tool
With only one tool, there is no possibility of ambiguity or overlap between tools. The tool's purpose is clearly defined as adding journal entries, and no other tools exist to cause confusion.
A single tool inherently has perfect naming consistency, as there are no other tools to compare it against. The name 'add_journal_entry' follows a clear verb_noun pattern (add + journal_entry).
A single tool for a journaling server is too few for the apparent scope. Journaling typically involves operations like viewing, updating, deleting, or searching entries, but this server only supports adding entries, which severely limits functionality.
The tool surface is severely incomplete for a journaling domain. It only allows adding entries, with no tools for retrieving, updating, deleting, or searching entries. This creates dead ends for agents, as they cannot perform basic CRUD operations beyond creation.
Maintenance
Related MCP Connectors
Plan trips directly into TravelOwl from a conversation with Claude.
First strength app Claude can write to: plan training in chat, it lands in the app ready to log.
Notes and actions in one app. Let Claude or ChatGPT read and update them.
- PenlogOAuthapp.penlog
Handwritten iPad journal your AI can read and write tasks back to, over MCP.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables structured note-taking with markdown support, dynamic tagging system, advanced search capabilities, and markdown export functionality through natural language conversations in Claude Desktop.3GPL 3.0
- FlicenseNot gradedqualityDmaintenanceConnects Claude Desktop to Obsidian vaults to enable reading, writing, searching, and intelligent organization of markdown notes. It features pre-configured structures for personal and family data management through natural conversation.2-
- AlicenseNot gradedqualityDmaintenanceEnables Claude to access and manage a persistent, human-readable knowledge graph of neurons, with semantic search, memory consolidation, and local ownership.MIT
- AlicenseAqualityCmaintenanceEnables personal knowledge management through Claude Desktop, allowing users to capture thoughts, connect ideas, and reflect on thinking changes via natural conversation.723MIT