A modern web application for tracking and managing your recurring subscriptions

White-box US macro research compiler: questions to real-data models, diagnostics, evidence graph, and auditable conclusions.

Paste your resume + a job description: get a local ATS/keyword fit score and AI-rewritten bullets. BYOK: OpenAI, Anthropic, Gemini, DeepSeek & more.

Describe your product -> AI generates a brand board: names, tagline, voice, a WCAG color palette, type pairing & an SVG wordmark. BYOK: OpenAI, Anthropic, Gemini, DeepSeek & more.

Turn any topic into a living, interactive AI knowledge map you can expand node by node. Bring your own API key: OpenAI, Anthropic, Gemini, DeepSeek & more.

Local-first BYOK portfolio analyzer for allocation, HHI, concentration, diversification, and AI-grounded explanations.

Local-first BYOK web app that computes 16 auditable financial ratios and uses AI only to explain the results.

Practice job interviews with an AI that adapts in real time and scores you per competency. Bring your own API key: OpenAI, Anthropic, Gemini, DeepSeek & more.

Upload a CSV + your API key; real in-browser analysis + an AI-written, chart-rich data story. BYOK: OpenAI, Anthropic, Gemini, DeepSeek & more.

Local web app and CLI that turns npm package-lock files into a dependency blast-radius map: paths, hotspots, duplicate versions, install scripts, and fixes

Local-first invoice intake and chase desk for freelancers: import CSV or pasted invoice text, review fields, prioritize overdue payments, and export follow-up plans.

Local-first return and warranty deadline desk for purchase exports, receipt text, action queues, and evidence checklists.

Local-first rental deposit evidence desk that matches deductions to proof, flags gaps, and drafts a dispute packet outline.

Open-source local-first workbench that reads your US medical bill, finds errors against the federal CMS taxonomy + No Surprises Act, drafts the appeal letter with deadlines.

Local Japanese pitch accent practice desk with browser recording, pitch-contour feedback, and mora-level drill exports.

把聊天记录画成K线图的 Skill。配合 relationship-candlestick-lab 项目使用，IDE 里的 LLM 按 v3.1 语义规则给每段对话打分，本地渲染成 K 线。

8-agent AI decision framework with deterministic Monte Carlo backend. LLM picks priors, Python runs the math.

A curated list of awesome Claude Skills, resources, and tools for customizing Claude AI workflows

Local-first circadian phase workbench: sleep log + chronotype + planned light -> multi-day timed light/dark protocol. Pure Python stdlib. Cited research.

个人昼夜节律相位工作台:输入 7 天睡眠日志 + 时型问卷 + 计划光照,生成多日逐小时光/暗协议。纯 Python 标准库,本地运行,文献可溯。

Local-first chat-log forensics workbench: K-S, M/M/1, Barabási heavy-tail, Granger. Pure Python stdlib. Cited research.

Multi-scale relationship-dynamics workbench: Lotka-Volterra ODE + SIR network + Gale-Shapley matching + curated dictionary. Pure Python stdlib. Local-first.

Wall Street IB-grade valuation workbench: DCF, P/E, P/B, EPS, sensitivity tables, forward-looking scenarios, archetype + strategy. Pure Python stdlib.

金枝 · The Golden Bough — 把聊天记录种成一棵世界树。看那些「如果当时」会长向哪里。

AI 决策引擎中文版 · 8 agent 协同 + 真实 Python 蒙特卡洛后端

把聊天记录画成K线图，看你和TA的关系曲线

Local-first chargeback evidence desk for small merchants: triage disputes, find evidence gaps, and draft response packets.

Local-first recurring charge audit desk that detects subscriptions, price creep, renewals, and cancellation priorities from statement exports.

Local-first inventory reorder radar for small shops using CSV exports.

Local trace-import lab for estimating AI agent workflow cost, retry risk, pricing sensitivity, and cost per successful run before launch.