Actually block video subtitles while learning a language, forcing yourself to listen
Problem
People want to train their ears with videos but can't resist reading subtitles; many players' subtitle-off switch doesn't work or only works on some platforms. There's no universal tool that truly covers the subtitle area.
Solution
A lightweight overlay tool that covers the subtitle region with an adjustable block or blur, working over browsers and local player windows, with a hotkey to peek and a mode that only applies when the mouse is away from subtitles—so you listen first, check later.
Track who paid for the shared subscription, until when, and who still owes money
Problem
People who've shared YouTube or Spotify subscriptions for years need to track not just what they pay for, but each member's paid-until date, uncollected payments, and how much they fronted—plus region switching, price hikes, proxies and gift cards. Existing subscription apps don't fit.
Solution
A subscription ledger built for group-sharing: track each seat member, billing cycle and due date, auto-compute uncollected and fronted amounts, support price-hike reminders, multi-currency and proxy payments, and export a reconciliation list.
Give Claude Code and other terminal coding agents a per-step shell command progress view
Problem
When a terminal coding agent runs shell commands, users only see a long spinner—no idea which step it's on, which command it's stuck at, or whether it already failed. The wait is opaque and failures require digging through logs.
Solution
A lightweight sidecar that reads the agent's shell command stream and shows each command's timing, status and output summary in real time, with failure highlighting and replay, making terminal coding observable.
Scan all your sites at once for known PHP/Node RCE exposure
Problem
When a PHP2Shell-style vulnerability lands, ops teams and small shops have a pile of sites and servers but no idea which versions are affected or which response headers leak the version. Existing tools only check a single endpoint, with no full-asset sweep. The exit of a global SSL/TLS certificate vendor also raises dependency-chain worries.
Solution
Input a list of domains or assets, asynchronously batch-check HTTP response headers, fingerprint PHP/Node versions, compare against known fixed releases, and output a report with severity and remediation advice—plus scheduled rescans and webhook alerts.
Turn high-value ChatGPT/Claude conversations into a searchable personal knowledge base in one click
Problem
People have valuable discussions with ChatGPT or Claude, then days later can't find them in the history list; manual copy-paste destroys code block and table formatting. It's a high-frequency chore repeated over and over.
Solution
A browser extension that exports whole sessions or selected messages from major AI platforms while preserving code blocks, tables and heading levels; outputs Markdown/JSON/PDF and can commit directly to the user's own GitHub/GitLab repo as a versioned knowledge base.
Browser-local TWR/XIRR with AI statement import—no data uploads
Problem
Individual investors want to know their true portfolio return (net of fees and taxes), which requires logging every transaction and computing TWR/XIRR. Existing tools often require accounts and data uploads, or only handle stocks and not mixed assets like rental property. Some end up building their own local tool.
Solution
A single-page web app storing data only in browser localStorage: users enter transactions manually or import them from PDFs/screenshots via AI, and it computes TWR and XIRR with gross/net-of-fee and after-tax views. It supports mixed asset classes including stocks, funds, and rental property, with report export.
Automatically log and quickly open every dev server you run locally
Problem
Developers often run multiple local services whose ports are scattered across terminals, and soon forget which service is on which port. Terminal history is lost after restart, forcing them to dig through project configs or restart. It gets especially messy when juggling several projects, making it hard to find a local page that was running earlier.
Solution
A small menu-bar/tray app that automatically watches new listening ports on your machine and records the associated process and directory. It shows a visual list where you can click to open in a browser or copy the address, with support for aliases, project grouping, and stale-port reminders.
One-command publishing of large multi-artifact projects to Maven Central with status polling
Problem
Publishing multi-module projects to Maven Central with many artifacts and large sizes (e.g., 90+ modules, 300MB+) repeatedly times out via the old OSSRH path, forcing manual re-submissions or full restarts. Developers must juggle staging, bundling, uploading, and polling, and switching between SNAPSHOT and release is error-prone.
Solution
A Gradle plugin/CLI that uses the Maven Central Publisher API to turn releases into a single command: aggregate all subprojects under the root, build and upload the bundle, and poll deployment status. It also supports direct SNAPSHOT publishing by auto-switching repositories based on the version suffix, reducing manual work.
One-stop self-check, version matching, and mitigation guidance for recent high-risk CVEs
Problem
Recent high-risk RCEs in PHP and Java (CVE-2026-109618/109619, CVE-2026-91403) force developers to juggle multiple projects, manually check versions, and hunt for mitigations—time-consuming and error-prone. Some spend a full day patching and are exhausted; others don't know if their stack is affected and dig through docs project by project.
Solution
A CVE-centric checklist tool: pick your stack (PHP/Java/Node, etc.) and deployment model to generate checkable self-audit steps; automatically compare versions against fix thresholds; provide mitigation config snippets for Nginx/Apache/WAF/Docker/K8s; export reports to sync with the team.
Automatically find and clean up hidden crash-dump bloat from Electron/Chromium apps
Problem
Many Electron/Chromium desktop apps (e.g., Codex desktop) keep writing Crashpad dump files locally when crash reporting hangs, silently accumulating tens to hundreds of GB and filling disks. Users only notice when the system slows down or runs out of space, and manual cleanup requires knowing to look in ~/Library/Application Support for Crashpad folders.
Solution
A lightweight menu-bar/tray tool that periodically scans Crashpad directories of known problematic apps (configurable patterns), visualizes disk usage, and offers one-click safe cleanup of old dumps. Set thresholds for alerts or auto-cleanup, plus a per-app disk-usage ranking to surface anomalies.
Make AI-generated pages follow a real design system and drop the AI smell
Problem
AI-generated pages often let the model improvise styling, producing inconsistent visuals that scream AI-made. People suggest feeding AI a mature design system and hand-writing copy to reduce the AI feel, but turning a design system into AI-usable specs and injecting it into the generation flow is still manual work.
Solution
Curate open-source design systems and extract them into tokens and component specs, offering a paste-ready prompt constraint pack plus a checker script. After generation, compare against the spec and suggest fixes for deviations.
Run regression tests and safety guardrails on prompts right in the browser
Problem
Before and after launch, a single prompt edit can skew outputs or leak sensitive content. Developers test ad hoc across scattered scripts with no unit-test-like regression, and while some browser-native prompt QA exists, most teams still lack a lightweight way to wire it into CI.
Solution
Maintain a suite of test cases in the browser (inputs, expected behavior, banned words, jailbreak instructions), run a one-click regression on each prompt change, and get pass rates, risk items, and diffs. Export results as CI-callable JSON and surface suggestions for blocking overreach instructions.
Patch many small servers by prompting an agent instead of SSH-ing into each one
Problem
Individuals and small teams running a dozen non-critical servers waste time patching and editing config host by host. Some already prompt Claude or Codex to operate the fleet and save hours, but lack a safe connector and worry an agent will delete files or misconfigure something.
Solution
A lightweight connector that, only on explicit approval, lets a local AI agent reach a declared host inventory: task templates for audits and patches, highlighted commands before execution, post-run logs, rollback scripts, plus directory and command boundaries.
Fix the recurring 'Repository not found' when one machine juggles multiple GitHub accounts
Problem
Developers with multiple GitHub accounts constantly hit 'Repository not found' on clone or push because an SSH key or credential matched the wrong account. Diagnosing means digging through ssh config, credential managers, and remote URLs, and it repeats with every new project.
Solution
A small CLI that inspects a repo's remotes, local SSH keys, and git credentials to determine which account is actually being used, checks it against the repo's expected account, and prints ready-to-run fix commands.
Real-time cost estimates for creators generating images and video through Claude and subscription tools
Problem
Creators using subscription image/video tools like Higgsfield face roughly a 50% markup over raw model pricing, unused credits expire every month even after paying, and they constantly copy-paste between Claude and the toolbox. People cannot tell what a single generation really costs, nor how many credits remain or when they vanish.
Solution
A lightweight pre-generation cost calculator: pull official model price tables, then apply the user's actual subscription billing rules to show true cost and markup multiplier. Track credits used, remaining, and expiry, and nudge users to spend expiring credits before they lapse.
Collect and organize community game port projects for easy discovery by game and platform
Problem
As native game porting barriers drop, a surge of community port projects appears, but information is scattered across forums and GitHub, making it hard for players to find usable projects by game or platform and know preparation steps.
Solution
Build an index site that catalogs community port projects by game and platform, showing project source, required preparation, and links to original projects, helping players quickly understand and participate.
Add controllable tone particles to AI replies for more human-like conversation
Problem
After chatting with AI for a while, users notice replies are correct but flat and overly polished, lacking human touch—like an instruction manual that never errs.
Solution
A pluggable expression layer that does not change answer content but adds tone particles to expression. Includes multiple Chinese styles (mature, cute, tsundere, etc.) and modes (fixed, weighted blend, random rotation), with density control and emoji toggle, while preserving code, tables, and numbers.
Quiz yourself on your own diff before git commit to verify you understand your changes
Problem
Developers often commit code they do not fully understand, making it hard to explain or debug later. Existing code review tools focus on style and errors, not the author's grasp of the change.
Solution
Run as a pre-commit hook, analyze the diff and generate a few targeted multiple-choice or short-answer questions; the developer must answer correctly to commit, otherwise they get hints to review relevant code or docs.
Check if your GPU or Mac can run a new model using real file sizes and KV cache
Problem
Every time a new open-source model drops, the community estimates VRAM by parameter count times bit width, which often diverges significantly from real requirements. Information is scattered and lacks credible sources.
Solution
Provide a checker per model that uses actual repository file sizes and computes KV cache layer by layer following llama.cpp allocation (including sliding window, linear attention, shared KV layers), then compares against available device VRAM/RAM. Each number is labeled by source type (official, community, estimate, measured).
An interactive CLI that cleans dependency and build junk for AI Agent developers
Problem
Agent development involves repeated dependency installs and builds, generating caches and artifacts that quickly fill up a 1TB disk. Generic cleaners cannot identify project types and risk deleting critical files.
Solution
Scan all projects in the workspace and report disk usage per project; detect common dependency and build output directories like node_modules, target, build; provide interactive confirmation before deletion to avoid removing source or config files.