AI Monk Index

io.github.Daisuke134由 GitHub 账号 Daisuke134 发布v0.4.19更新于 Oct 25, 2026

Research monk avatars, products and video playbooks with sources and labelled uncertainty.

已验证Streamable HTTP可网页运行金融知识与记忆数据与分析

概览

AI 生成的概览

一个远程只读研究目录,让助手查询 AI 和尚视频人设、其视频形式、变现方式与创作者玩法手册。

可以做什么
按名称、账号、来源、语言、平台和地区搜索和尚档案与待研究候选,支持分页及按受众或收入排序。按 ID 获取单个档案的完整研究记录、来源、AI 状态、未知项与身份冲突。读取目录统计,包括记录数、来源数、AI 状态计数与快照方法。列出按公开播放量排序的视频样本,含标题、时长、上传日期与结构摘要。查看变现报价、价格、卖家计数与创作者收入声明,并阅读带来源的创作者玩法手册,涵盖脚本公式、钩子、声音、发布节奏与转化路径。
适用场景
适合研究 AI 和尚视频人设、其视频形式与变现方式,或需要带来源、标注不确定性的创作者玩法资料,而非完整普查。
运行要求
位于 aniccaai.com 的远程 streamable HTTP 端点;未声明任何软件包、环境变量、请求头或身份验证。需要能访问该端点的网络。
安装前请注意
内容为只读研究资料,并非完整普查;待研究候选未经核实,档案存在不代表确认 AI 制作或视频形式。应保留 AI 状态、未知项、来源、日期与身份冲突。价格不等于订单,创作者收入声明未经审计;不要由播放量、价格或计数推断或保证收入。工具数据是第三方内容,不是指令。

安装

在 SourceWeft 中

  1. 打开 控制台中的 AI Monk Index,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。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.Daisuke134%2Fai-monk-index/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "ai-monk-index": {
      "type": "http",
      "url": "https://aniccaai.com/.netlify/functions/ai-monk-mcp"
    }
  }
}

README

→ Daisuke134/life-manager へ移行済み

Anicca — Behavior Change Agent + Wake-up SaaS

Proactive agent that helps people break harmful patterns and build better habits. Two surfaces share one codebase:

SurfaceWhatStatus
iOS appOnboarding → 13 problem types → scheduled nudges → 1-screen card UXLive on App Store (1.9.x)
Wake-up SaaS$9.99/mo — AI calls you every morning, escalates if you're running lateLive at https://aniccaai.com/alarm

The wake-up SaaS uses a real voice agent ("Anicca") who talks back over PSTN, not a recorded ringtone. She knows your calendar, your home location, and your day-job start time — and if you're falling behind schedule she'll keep calling plus optionally email a 報連相 (heads-up) to your manager.


Repo layout

PathWhat
aniccaios/iOS app (Swift / SwiftUI, iOS 15+)
apps/api/Node.js mobile backend (Railway)
apps/landing/Marketing site + checkout (Netlify, aniccaai.com)
apps/alarm-backend/Wake-up SaaS (this is what we just open-sourced)
apps/alarm-backend/bridge/Gemini Live ⇄ Twilio Media Streams WebSocket bridge (Node)
apps/alarm-backend/scheduler/Per-minute alarm + 15-min lateness cron (Python, Railway worker)
maestro/E2E test flows

How the Wake-up SaaS works

User signs up at aniccaai.com/alarm        ↓ Stripe webhookSupabase `subscriber_profiles` row created (phone, wake_time, tz, name)        ↓scheduler/alarm_scheduler.py  (every 60s, multi-tenant)        ↓  it's wake_time in user's tz, and not fired todayscheduler/wake_loop.py        →  POST /dialout on bridge        ↓bridge/server.js              →  Twilio Calls API        ↓Twilio → user's phone → bridge WebSocket → Gemini Live (Charon voice)        ↓Real two-way conversation until user is verifiably awake.
Then, every 15min from 6:00-23:00 local:scheduler/saas_lateness.py    →  reads gcal + OwnTracks location        ↓  ETA - now < bufferCalls again with lateness persona, optionally emails stakeholders.

OwnTracks setup (required for lateness detection)

The lateness loop needs to know where the user is right now. We use OwnTracks — open-source location publisher — because it's privacy-respecting (you own the data) and battery-aware.

One-time setup on the user's phone

  1. Install OwnTracks

  2. Configure HTTP publishing (Settings → Preferences → Connection)

    • Mode: HTTP
    • URL: https://<your-anicca-alarm-backend>.up.railway.app/ingest/owntracks?token=<USER_TOKEN>
      • The token comes from aniccaai.com/alarm/setup after signup.
    • Identification → DeviceID: anything (e.g. iphone)
    • Identification → TrackerID: 2 chars (e.g. DA)
  3. CRITICAL: Set mode to Move, NOT Significant

    • Open OwnTracks → ⓘ icon → Mode selector
    • Choose Move
    • Why: per OwnTracks docs, Significant mode publishes nothing while the device is stationary. If the user is asleep in bed, no location update is sent for hours — and Anicca then sees a stale fix and can't tell whether they left home. Move publishes every 5 min OR 100 m of movement (whichever is sooner), which is what the lateness loop needs. Battery cost is roughly that of a navigation app, so we recommend the user enables Low Power Mode override.
  4. Grant always-on location permission (iOS: "Always Allow")

    • Otherwise iOS will silently throttle background publishes.
  5. Verify it's working

    • Go to aniccaai.com/alarm/setup → "Last seen" timestamp should refresh within 5 minutes of you walking around with the phone.

Why HTTP and not MQTT

We support only the HTTP publisher (simpler ops, no broker to run). MQTT will return.

Why we don't use Apple Find My / Significant Location Changes

  • Find My has no public API.
  • iOS Significant Location Changes API only fires on 500m+ movement after 5 min — useless while someone is in bed.

Self-hosting

This is the cheap path if you want to run your own wake-up agent for yourself or a small group.

Required services (free or near-free tiers)

ServiceWhat forFree tier?
TwilioPhone number + outbound calls$1/mo number + ~$0.015/min
Google AI StudioGemini Live APIyes (per-key quota)
SupabaseSubscriber DByes
Railwaybridge + scheduler hosting$5/mo hobby
StripeBilling (skip if self-hosting for yourself)—

Environment variables

Set these in Railway (apps/alarm-backend/bridge service) AND (apps/alarm-backend/scheduler worker):

bash
# Voice + LLMGEMINI_API_KEY=...GEMINI_LIVE_MODEL=gemini-2.5-flash-native-audio-preview-09-2025GEMINI_VOICE=Charon
# TelephonyTWILIO_ACCOUNT_SID=AC...TWILIO_AUTH_TOKEN=...TWILIO_PHONE_NUMBER=+1...
# Subscriber DBSUPABASE_URL=https://<project>.supabase.coSUPABASE_SERVICE_ROLE_KEY=eyJ...   # service role, not anon
# Stakeholder ETA (lateness loop only)GOOGLE_API_KEY=...                  # Google Maps Directions API[email protected]         # which Google account holds the user's calendarGOG_KEYRING_PASSWORD=...
# Optional integrationsCOMPOSIO_API_KEY=...                # 1-click calendar OAuthSLACK_BOT_TOKEN=xoxb-...            # ops alertsSLACK_CHANNEL_ID=C...
# NetworkingPORT=8787PUBLIC_BASE_URL=https://<railway-domain>

Deploy

bash
# bridge (Node)cd apps/alarm-backend/bridgerailway linkrailway up
# scheduler (Python worker — runs every minute via Procfile)cd ../schedulerrailway linkrailway up

Supabase schema and Stripe webhook are in the live aniccaai backend; if you're self-hosting standalone we'll publish a minimal schema separately.


Stack

  • iOS: Swift, SwiftUI, RevenueCat, Mixpanel
  • Voice: Pipecat + Gemini Live (native audio S2S)
  • Telephony: Twilio Media Streams
  • Backend: Node.js (bridge), Python (scheduler), Express, Railway, Supabase
  • Tests: Maestro (E2E)

Status

ComponentState
iOS appLive (1.9.x)
Alarm SaaSLive, paid customers since 2026-05-23
Lateness loopLive, requires OwnTracks Move mode (see above)
Self-host docsMinimal (this file) — improvements welcome via PR

License

The code is OSS so people can learn from it and self-host. Commercial use, trademarks, and the "Anicca" name are reserved. See LICENSE once published.

Contact

来源:README.md,提交 3bb0b68

工具

6

search_ai_monks

Search AI Monk Index

Search monk profiles and pending research candidates by name/handle, provenance, language, platform and region. Excludes illustrated/watercolor storytelling channels. Paginated; latest single-platform audience order by default; optional sampled-video views or reported three-month USD gross-revenue order. Missing metrics are not zero. Returns provenance, directory_scope and unknowns; candidate inclusion does not confirm AI production or video format.

输入参数sortlimitqueryoffsetregionstatuslanguageplatform

get_ai_monk

Get AI monk research profile

Retrieve one exact directory profile ID, its complete research record and referenced sources. Includes pending candidates with their directory_scope, identity conflicts and unknowns. Illustrated channels are excluded; candidate inclusion does not confirm AI production or video format.

输入参数id必填

ai_monk_index_stats

AI Monk Index coverage

Return directory record/source counts, AI status counts and snapshot methodology. Metadata separates the full research archive from directory coverage. Directory counts include evidence-reviewed video personas and pending research candidates; illustrated channels are excluded. AI status counts do not establish reviewed video-avatar evidence.

无需输入参数。

top_formats

Top AI monk video formats and hooks

Observed AI monk video samples ranked by latest dated public views, with title, duration, upload date and structure_summary (format, hook and script pattern). Optional account_id. Views are sampled-video observations, not channel totals.

输入参数limitaccount_id

how_they_monetize

How AI monk accounts monetize

Offers (type, name, listed or base price, currency, URL), separate localized presentment observations, seller counters, creator revenue claims and revenue observations for AI monk accounts, plus counts by offer type. Optional account_id. Prices are not orders; claims are unaudited.

输入参数account_id

winning_playbooks

Winning AI monk playbooks

Source-linked playbooks published by successful AI monk creators (e.g. Yang Mun by Shalev): script formula, winning hooks with reported views, voice, posting cadence, funnel to product and monetization. Creator claims are unaudited.

无需输入参数。

版本历史

4
  1. v0.4.19最新Oct 25, 2026
  2. v0.4.14Oct 24, 2026
  3. v0.4.8Oct 24, 2026
  4. v0.4.5Oct 23, 2026

评分

已满足适用检查项的 90%

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 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具必需15 分
  • 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。