ai-mood-journal

An MCP (Model Context Protocol) server for children's daily mood check-ins, emotional pattern tracking, and weekly parent digest reports — all stored locally with end-to-end encryption.

Motivation

Adolescent mental health is a growing concern. Research consistently shows that early detection of mood changes, coupled with warm parental engagement, significantly improves outcomes for children and teens. This tool gives parents a gentle, privacy-preserving way to stay connected to their child's emotional wellbeing — without surveillance or pressure.

The design is intentionally non-clinical: check-ins feel like a natural part of AI conversations, not a therapy session. Data never leaves the local machine. Parents receive warm, actionable summaries — not raw transcripts.

Installation

1. Install dependencies

npm install

2. Configure environment variables

Copy .env.example to .env and fill in the values:

cp .env.example .env

Required:

• MOOD_JOURNAL_DB — path to SQLite database file (default: ~/.ai-mood-journal/journal.sqlite)
• MOOD_JOURNAL_ENCRYPT_KEY — 32-byte encryption key (hex or base64); generate with:

  node -e "console.log(require('crypto').randomBytes(32).toString('hex'))"

Optional (for notifications):

• Telegram: TELEGRAM_BOT_TOKEN, TELEGRAM_PARENT_CHAT_ID
• Email: SMTP_HOST, SMTP_PORT, SMTP_USER, SMTP_PASS, SMTP_FROM, SMTP_TO

3. Build

npm run build

4. Configure Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "ai-mood-journal": {
      "command": "node",
      "args": ["/path/to/ai-mood-journal/dist/index.js"],
      "env": {
        "MOOD_JOURNAL_PROFILE_ID": "alex",
        "MOOD_JOURNAL_CHILD_NAME": "Alex",
        "MOOD_JOURNAL_CHILD_AGE": "13",
        "MOOD_JOURNAL_DB": "~/.ai-mood-journal/journal.sqlite",
        "MOOD_JOURNAL_ENCRYPT_KEY": "your-32-byte-hex-key",
        "TELEGRAM_BOT_TOKEN": "optional",
        "TELEGRAM_PARENT_CHAT_ID": "optional"
      }
    }
  }
}

Or using npx after publishing:

{
  "mcpServers": {
    "ai-mood-journal": {
      "command": "npx",
      "args": ["-y", "@gonzih/ai-mood-journal"],
      "env": { "...": "..." }
    }
  }
}

MCP Tools

check_in

Log a child's daily mood check-in. One entry per day per profile (upserts on repeat calls).

Input:

{
  "profileId": "alex",
  "mood": "okay",
  "freeText": "Had a rough day at school but practice was fun."
}

Mood values: great (5) | okay (3) | not_great (2) | frustrated (2) | anxious (2) | tired (2) | numeric 1-5

Output:

{
  "logged": true,
  "streakDays": 5,
  "encouragement": "Thanks for sharing! Remember, okay days are perfectly fine. 😊",
  "flags": []
}

Automatically detects crisis language and sends an immediate alert if found.

get_mood_summary

Get mood trends and concern flags over a rolling window.

Input:

{ "profileId": "alex", "days": 7 }

Output:

{
  "avgMood": 3.2,
  "trend": "stable",
  "flags": [],
  "topTopics": ["school", "friends"],
  "streakDays": 5
}

Trend values: improving | declining | volatile | stable

get_weekly_digest

Generate a warm, narrative weekly summary for parents.

Input:

{ "profileId": "alex", "weekStart": "2026-03-15" }

Output:

{
  "digest": "📊 Alex's Week in Review — Mar 15 – Mar 21\n\n...",
  "flags": []
}

The digest is cached in the database for the given week.

list_entries

Retrieve recent mood entries (free text is decrypted before returning).

Input:

{ "profileId": "alex", "limit": 30 }

Output: Array of mood entry objects with decrypted free_text, parsed topics array, mood scores, and dates.

send_digest

Generate the digest for the current week and send it via Telegram (preferred) or email (fallback). Updates sent_at in the database.

Input:

{ "profileId": "alex" }

Output:

{ "sent": true, "channel": "telegram" }

get_guidance

Get parent-facing guidance and resources for a specific concern flag.

Input:

{ "flag": "persistent_anxiety" }

Output:

{
  "guidance": "Anxiety appearing frequently. Ask: 'What's feeling heavy right now?'...",
  "resources": ["childmind.org/article/anxiety-in-children", "anxietycanada.com/resources"]
}

Flag values: consecutive_low_mood | declining_trend | acute_drop | persistent_anxiety | social_withdrawal | crisis_language

Privacy

• Local-only storage: All data lives in a SQLite file on your machine. Nothing is sent to any cloud service.
• Encrypted free text: Children's reflective writing is encrypted with AES-256-GCM before being written to disk. The key never leaves your environment.
• Theme-only digests: Parent reports surface topics and mood trends — never verbatim child quotes.
• Optional notifications: Telegram/email are opt-in. If not configured, alerts are only visible in the tool output.

Crisis Protocol

If a child's free text contains language suggesting self-harm or suicidal ideation, the server:

1. Immediately fires sendAlert() with a crisis-prefixed message to the configured Telegram/email channel.
2. Returns 'crisis_language' in the flags array of the check_in response.

The AI layer (per SKILL.md) should halt the normal session and respond with warm, direct encouragement to involve a trusted adult.

Crisis resources always included in get_guidance('crisis_language'):

• 988 Suicide & Crisis Lifeline: call or text 988 (US)
• Crisis Text Line: text HOME to 741741
• NIMH Suicide Prevention: nimh.nih.gov/health/topics/suicide-prevention

Configuration Reference

| Variable | Required | Description | |---|---|---| | MOOD_JOURNAL_PROFILE_ID | No | Default profile ID | | MOOD_JOURNAL_CHILD_NAME | No | Child's name for digest reports | | MOOD_JOURNAL_CHILD_AGE | No | Child's age | | MOOD_JOURNAL_DB | No | Path to SQLite file (default: ~/.ai-mood-journal/journal.sqlite) | | MOOD_JOURNAL_ENCRYPT_KEY | Recommended | 32-byte key as hex (64 chars) or base64 | | TELEGRAM_BOT_TOKEN | No | Telegram bot token for notifications | | TELEGRAM_PARENT_CHAT_ID | No | Telegram chat ID for parent notifications | | SMTP_HOST | No | SMTP server hostname | | SMTP_PORT | No | SMTP port (default: 587) | | SMTP_USER | No | SMTP username | | SMTP_PASS | No | SMTP password | | SMTP_FROM | No | Sender address | | SMTP_TO | No | Recipient address | | DIGEST_DAY | No | Day for weekly digest (default: sunday) | | DIGEST_TIME | No | Time for weekly digest (default: 18:00) |

Research Foundation

This tool draws on established research in adolescent psychology and mood tracking:

• Beck, A.T. (1979). Cognitive Therapy of Depression. Guilford Press.
• Hankin, B.L., & Abramson, L.Y. (2001). Development of gender differences in depression: An elaborated cognitive vulnerability–transactional stress theory. Psychological Bulletin, 127(6), 773–796.
• Kovacs, M. (1992). Children's Depression Inventory. Multi-Health Systems.
• Lewinsohn, P.M., Hops, H., Roberts, R.E., Seeley, J.R., & Andrews, J.A. (1994). Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-III-R disorders in high school students. Journal of Abnormal Psychology, 102(1), 110–120.
• Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology, 100(4), 569–582.
• Silk, J.S., Steinberg, L., & Morris, A.S. (2003). Adolescents' emotion regulation in daily life: Links to depressive symptoms and problem behavior. Journal of Youth and Adolescence, 32(6), 365–376.

License

MIT © Gonzih