Log every shared expense and get a minimal transfer path when the trip ends.
Problem
After trips, dinners, or group activities, reconciling in chat apps is painful: who fronted the money, who still owes, and a pile of small transfers going back and forth. One developer built a mini app for personal use and it accumulated over 900 group ledgers, proving demand is real.
Solution
A lightweight tool focused on multi-person splitting plus optimal transfer paths. Each entry assigns participants, and when the activity closes it computes the fewest transfers and exports a clear result page. Data is immutable and visible to all members, reducing trust friction.
Scan local coding agent configs and skills to reveal what is quietly burning tokens.
Problem
More people complain their quota runs out but cannot pinpoint which SKILL, MCP config, or memory file is eating tokens. Some tried RTK truncation only to be told key info was dropped, leaving them to manually tweak CLAUDE.md, skills, MCP, and hooks by gut feel.
Solution
A local desktop tool that reads each coding agent’s config and skill directories, estimates fixed context cost per skill and MCP, and charts token usage per session. It surfaces concrete advice like disabling a skill, switching to manual trigger, or using an index, with one-click backup and config switching.
Paste Windows screenshots directly into WSL, terminals, and remote Codex sessions.
Problem
Developers running Codex or terminal agents inside WSL hit “no image on clipboard” when trying to paste a screenshot, because the WSL clipboard does not bridge to the Windows graphical clipboard. Some switch back to desktop apps just to paste, which breaks the workflow.
Solution
A small Windows tray app watches clipboard images, syncs them into files accessible from WSL, and exposes a CLI plus hotkey for one-tap use from the WSL side. It can also act as a paste hook for VS Code or terminals, converting images to a path or base64.
A quota alert and fallback router for heavy AI coding users so a 429 never blocks a deadline again.
Problem
Heavy users run multiple coding projects across Codex and Claude Code, then a model returns “Selected model is at capacity” or endless 429s and the finishing work just stalls. There is no unified quota view across accounts and models, and reset notifications rely on ad-hoc community tools.
Solution
A local proxy plus menu bar app that routes requests from Claude Code, Codex, and Cursor through one layer. It shows live quota and 429 status per model, automatically falls back to an alternate model or queues retries on limit hit, and pushes reset notifications through user-configured channels.
Gather all guest photos into one place automatically
Problem
After a wedding or event, photos are scattered across texts, Instagram, Google Photos, and everyone's phone. Couples struggle to collect them all. Existing sharing is fragmented and guests resist downloading new apps.
Solution
Create an event album link; guests upload photos and videos directly via web without installing an app. Supports QR code join, auto-categorization by uploader, shared album, and optional AI curation.
Many small business websites look nice but visitors can't tell what the business actually does, the homepage doesn't build trust, and there's no clear next step. Owners struggle to spot these issues themselves.
Solution
Enter a website URL, automatically analyze above-the-fold copy, value proposition clarity, trust signals (reviews, case studies, credentials), and conversion path, then generate an actionable improvement checklist with priority ranking.
Check if your product name is already taken before launch
Problem
Founders spend months building under a name, only to discover four competitors already use it—one launched just six weeks earlier. Existing tools only check trademarks or domains, not actual product footprints on app stores, Product Hunt, or GitHub.
Solution
Enter a product name, scan App Store, Google Play, Product Hunt, GitHub, Crunchbase and more for identical or similar product listings, return a conflict severity score and recommendations.
Keep Gmail, let AI organize, summarize, and nudge replies
Problem
Important emails get buried, forcing users to manually search, flag, and reply. Some want an AI email manager but fear replacing Gmail or privacy issues. The core pain: don't want to switch email, just need an assistant that continuously understands the inbox and nudges action.
Solution
Using Gmail API read-only access, continuously analyze the inbox, automatically categorize important emails, generate summaries, detect items needing reply or action, and push reminders. All data processed locally or encrypted, no email body stored.
A real task graph for people blocked by prerequisites
Problem
Many people are labeled procrastinators when they are actually blocked by real dependencies. For example, someone wants to clear out a garage but must wait for two different people to remove heavy equipment first. Existing to-do lists show 'incomplete' without showing 'why it can't be done.' This causes frustration and makes it hard to explain progress to others.
Solution
Input tasks and prerequisites, then automatically generate an interactive dependency graph that highlights only truly actionable tasks and shows who or what is blocking them. Supports reminders and notes so others can see where you're stuck.
One persistent memory shared across different AI coding agents
Problem
AI coding agents lack persistent memory, so switching tools or sessions means re-explaining project context every time. Someone showed a self-hosted persistent memory layer via MCP for AI coding agents, proving this is a real, actively addressed pain point.
Solution
A local-first memory service compatible with MCP that imports session logs from different agents, deduplicates them into a project-level memory store, and automatically injects relevant context into new sessions.
Drag your recent Mac screenshots anywhere, instantly
Problem
Apple's screenshot experience is bad: you take a screenshot and then race to drag it into the target app, or dig through folders. Someone built a simple extension where a shortcut opens a drawer of recent screenshots that are fully drag-and-droppable.
Solution
A minimal macOS menu bar app that watches the screenshot folder, opens a recent-screenshot panel with a global shortcut, supports search, pinning, dragging into any app, and one-click path copy.
Get alerted before bot traffic explodes your cloud bill
Problem
A developer's Vercel bill jumped 10x not from user growth but from over 1M requests by meta-externalagent hitting a preview deployment. They never worried about bot traffic before and had no alerting, so they only found out when the bill arrived.
Solution
A lightweight monitoring service that connects to platform event logs, tracks bot traffic share and request spikes, alerts via email or chat before costs spike, and suggests blocking rules.
One-click benchmark of how fast your Mac runs local LLMs
Problem
Developers want to know real throughput and memory usage of local models on their own machines, but only have scattered personal anecdotes. Someone asked the community for the best local model on Mac with decent speed and posted their own M5 Pro 48GB numbers (Qwen 3.8 Flash at 14-15 tok/s), showing a clear lack of standardized tooling.
Solution
A local benchmark tool that auto-detects hardware, points at user models, runs standard prompts, measures throughput and memory, and generates shareable comparison cards and leaderboards.
One-click migration from existing fail2ban configs to lightweight single-binary alternatives
Problem
Ops engineers like fail2ban's log-to-ban model but dislike maintaining a Python runtime and regex attack surface. Someone already built fail2zig and specifically added config inspection and migration from fail2ban, proving there's real demand for migrating existing setups.
Solution
A cross-platform CLI/GUI that reads existing fail2ban jails, filters, and actions, parses and converts them into the target format, flags incompatible regex with suggestions, and produces a migration report plus rollback bundle.
Play phone-shot vertical videos side by side on TV and kill the black bars
Problem
A developer first built a tool to download Douyin and other platform clips into iCloud Photos, then realised vertical videos look terrible on TV with huge black bars on both sides. The fix was playing three at once, filling the big screen and letting the remote switch audio focus between left, center and right. A group of users already has local vertical libraries and no player designed for them.
Solution
A TV player that splits the screen into three columns playing three vertical videos in sync, with one-touch audio focus switching on the remote, plus random or single-album playback from Photos or a NAS. An iPad companion offers two-up side-by-side mode for workouts or tutorials, all reading from a unified local media index with no extra uploads.
A lightweight chat client with no analytics, no cloud, and your own key
Problem
A developer noticed that after ChatGPT started pushing ads, their Instagram reels became increasingly aligned with what they had talked about in ChatGPT, and decided to quit it for good by building a browser-based client with no backend. What people want is not a stronger model but an interface where they can verify where their data actually goes.
Solution
A pure front-end single-page app where users plug in their own API key or point to a local model, and all conversations stay in browser IndexedDB. No account, no analytics script, no server. It ships one-click wipe, encrypted session export, multi-model switching and a public data-flow document so anyone can audit it.
Slice long Claude Code sessions into readable phases so you can see where it went off the rails
Problem
A developer spent weeks trying to write a simple paper with Claude and, despite heavy context engineering and material prep, the output degraded into an unintelligible word salad. Meanwhile the community keeps building tools to make AI output understandable because nobody wants to read the entire long tail of decisions and garbage. The problem is not model strength — it is that once a session grows long, humans lose any sense of when it started drifting.
Solution
Parse local Claude Code session logs, automatically split them into phases based on goal switches, tool-call bursts, retry storms and context trimming, and show duration, token burn and success rate per phase. A timeline highlights likely drift points, and any phase can be exported for review or a bug report.
Use your own model and prompts to process documents inside Word, Excel and PowerPoint
Problem
Developers are tired of the black-box, locked-in AI inside MS Office: one vendor decides which models you can use, where document data goes, and what the agent is allowed to do, with no source to inspect and no way to PR a feature. Many people are forced to use Office at work and simply have no choice. Coding agents already let you bring your own model — office docs have not caught up.
Solution
An Office add-in that opens up the AI layer: users plug in any OpenAI-compatible endpoint or local model, inspect and edit the prompt behind every action, and see exactly which model received which data. The core loop is select text, run an auditable operation, and roll it back in one click.
Help wrongly suspended small businesses assemble a structured appeal and migration plan in minutes
Problem
A business running for almost 15 years had its Google Business Profile suspended after simply updating its address, and the appeals process was rejected with no clear path forward, costing it discoverability and customer reviews. Meanwhile other users report that even with correct password, MFA and security keys, Google account login fails with "we couldn't verify it's you" for days. Platforms give you a blank text box, and small operators have no template, no evidence checklist, and no idea what reviewers actually look for.
Solution
An offline-first tool that guides the user through collecting account details, business history, address-change proof and review screenshots into a clean timeline, then generates platform-aware appeal letters in English and Chinese with a step-by-step submission and follow-up plan. It also ships a migration checklist so the owner can move customers and traffic to channels they control while locked out.
Remap the Chrome shortcuts the browser refuses to let you change
Problem
Many users have muscle memory for tab or window control, but Chrome's built-in shortcut customization is weak. Someone wanted Command+Shift+Up/Down to switch tabs, searched for extensions, and none could actually remap the keys—so they built their own. That suggests key bindings that extensions often cannot touch, and the need is not unique.
Solution
A focused extension that lets users remap actions Chrome normally won't: tab switching, window switching, address-bar actions and more, no dev background needed. It includes conflict detection, preset templates, install-and-go behavior, and config import/export. It is explicitly a patch layer for what the browser does not do.