One-click connect two computers for encrypted P2P file transfer, intranet penetration, and remote SSH.
Problem
Existing file transfer tools either rely on cloud storage (like the discontinued 奶牛快传) or require complex setup (like native Tailscale CLI). Users need a simple, clean, out-of-the-box P2P solution.
Solution
A cross-platform desktop app based on the open-source tailcat project, providing a GUI. Open the app to directly connect two computers, supporting encrypted P2P file transfer, intranet penetration, SSH, chat, and audio/video without public IP or complex config.
Edit documents with natural language directly in browser using local small language models, no internet or API costs.
Problem
Online document editors typically send content to cloud AI services, incurring API costs and privacy concerns. Users want to operate documents with natural language without internet or fees.
Solution
A browser-based document editor built on OnlyOffice + WASM, integrated with WebLLM and local small models. Supports natural language commands (select, rewrite, summarize, comment, format, export) all processed locally, no API calls, completely free.
Restore old photos online for free 3 times daily, no signup, no watermark, no charge for failures.
Problem
Most old photo restoration tools hide behind subscription walls after processing, charge for failed tasks, watermark free results, and users worry about images being used for training.
Solution
A minimalist online tool that restores, upscales, and sharpens photos upon upload. No login, 3 free uses per day, no watermark (free version max 2048px), transparent credit display before processing, automatic refund for failures, and images auto-deleted within 24 hours. Paid via one-time credit packs that never expire.
Filters noise for founders and indie hackers, delivering only high-quality projects and monetization cases.
Problem
Indie hackers and global founders face severe information overload: Product Hunt, Hacker News, and X are flooded with wrapper tools, crypto ads, and PR fluff. Finding genuinely useful hardcore projects or real MRR cases is like finding a needle in a haystack.
Solution
A website aggregating first-hand sources like Product Hunt, HN Show HN, Trends.vc, and V2EX Creative, using LLMs and semantic clustering to filter low-quality content and surface verified tools, project updates, and monetization cases. Supports email or Telegram push.
Warns you before you accidentally send personal identifiable information online.
Problem
Users often unknowingly leak PII like phone numbers, ID numbers, or addresses when filling forms, sending emails, or chatting. Existing tools mostly do static scanning and can't prevent real-time leaks.
Solution
A lightweight browser extension that scans input fields and editable text in real-time before submission, detects common PII patterns (phone, email, ID, bank card), and warns users with an option to redact. Supports custom sensitive words and regex rules.
See which processes and which Wi-Fi ate your hotspot data this month
Problem
Remote development over a phone hotspot can burn nearly 10 GB a day, but the carrier only tells you the total, not where it went — remote desktop, dev tools, browser, or background downloads? macOS built-in tools can't attribute traffic clearly by process or by Wi-Fi network name, leaving users guessing.
Solution
A ~1MB macOS menu-bar app showing today / yesterday / this-month total up and down traffic, per-process rankings with search and sort, network differentiation by real Wi-Fi name (e.g. phone hotspot), and live speed in the menu bar. Fully local, no server required.
Structure exploratory AI conversations into goals, constraints, open questions, and acceptance examples
Problem
Discussing requirements with AI starts fuzzy: constraints emerge, earlier judgments get revised, and open questions pile up. Afterward it's hard to say what was actually settled, what's confirmed, and what's still exploratory. Asking the AI to summarize flattens disagreements and uncertainty, leaving downstream tasks on shaky ground.
Solution
A conversation-distillation tool that organizes imported chats or manual excerpts into layers — task goals & constraints / confirmed judgments / supporting rationale / open questions / revised prior beliefs / acceptance examples — each with a confirmation status. The structured export feeds directly into coding agents as task constraints.
One observability panel for all local AI sessions and tool calls, so vibe-coding isn't a black box
Problem
More people run multiple parallel tasks in Codex / Claude Code, yet have zero visibility into whether processes are stuck, how much they consume, or what each session produced. Some get rate-limited or suspect they're being throttled, and others complain that big-vendor AI tools are buggy and rough. The missing piece isn't another model — it's visibility into the local AI workflow.
Solution
A menu-bar / tray dashboard that auto-discovers locally running AI tool processes and sessions, showing status, token usage, duration, and recent output per session, with alerts for stalls or quota anomalies. All data stays local.
A self-hosted tool that auto-clears your GitHub notifications by rules
Problem
GitHub notifications pile up to nearly 100 a day, and the built-in filters are so weak that even marking merged PRs as Done takes multiple pages. Many devs start every day by clearing notifications — a frustrating, high-frequency chore. Someone wrote a rule-based auto-cleaner specifically because of this, confirming the pain is real and tolerated.
Solution
A small self-hostable service: pull notifications via the GitHub API, apply nested condition rules (repo, type, state, keywords, merged status, etc.) to auto-mark Done / mute / unsubscribe, and preview which notifications a rule would clear before applying. Rules are exportable as JSON for teams.
Turn any webhook POST into a loud phone alert for self-hosted services, Jenkins, and Uptime Kuma
Problem
Ops folks and self-hosters still rely on email or Slack for alerts, which are easy to miss and not loud enough on a phone. They want to POST to a webhook and get a phone push, but existing options either require running a server or wiring up multiple platforms. An iOS app was built precisely because midnight alerts weren't visible enough, showing the pain is real and long tolerated.
Solution
A minimal webhook-to-phone-push relay: send an HTTPS POST to a dedicated endpoint and get a high-priority push instantly, with in-app acknowledge / mute / do-not-disturb and pairing codes to bind a second device. Ship templates for Jenkins, Cloudflare, Uptime Kuma, and GitHub Actions so alerts work out of the box.
Zero-server private Git hosting on Cloudflare, auto-syncing your GitHub
Problem
Some repos you don't want on GitHub, or you want off-site backups of your private GitHub repos, but self-hosting GitLab or Gitea requires a server and ongoing ops—too heavy for individuals and small teams.
Solution
A ready-to-deploy setup that hosts Git repositories on Cloudflare Workers, D1 and R2 with no server to maintain. It includes a GitHub App that automatically syncs all repos from authorized accounts. Users click deploy and they're done.
Cut through AI noise and find authors and podcasts actually worth following
Problem
Software practices are changing faster than ever with AI, and blogs, podcasts and Twitter produce overwhelming volume, most of it noise. Developers know authors like Simon Willison are worth following, but have no systematic way to discover similar trusted sources beyond word of mouth.
Solution
A human-plus-algorithm maintained directory of developer sources: by topic (AI coding practices, agent architecture, code review, etc.) it lists community-vetted authors, podcasts, and newsletters, with recent activity and a one-line reason to follow. Users can subscribe to updates and submit or upvote recommendations.
Watermark your ID and passport photos without uploading anything
Problem
Renting an apartment, opening an account, or applying for a visa often requires uploading a copy of your ID. Handing over the original image exposes your identity, and most online watermark tools require uploading the file, which defeats the purpose.
Solution
A purely client-side web tool: drag in an image and generate a version with custom text (e.g. 'for X use only'), date, and diagonal pattern, all processed locally. Neither the original nor the watermarked image leaves the browser. Supports batch processing, PDF export, and optional redaction of sensitive fields.
A task manager for all your running AI coding sessions
Problem
Many developers run several Claude Code or Cursor agent sessions at once and quickly lose track of which is doing what, which is stuck, and which is burning credits. The terminal becomes a mess with no process view, forcing reliance on memory and window shuffling.
Solution
A local menubar/tray app that auto-discovers all active agent sessions and shows status (running, awaiting input, done, possibly stuck), model credit usage, and a summary of recent output. One click to interrupt, jump back to the terminal, or set a completion alert for long jobs.
Let AI agents run their generated code locally without burning your machine
Problem
Developers increasingly let coding agents write and execute scripts directly, but almost nobody dares run agent output on a real machine. Rolling your own sandbox is either dependency-heavy or has no real isolation. Treating agent-generated code as untrusted input has become a baseline requirement.
Solution
A self-hosted lightweight sandbox where agent-generated code runs first under strict policy: network, filesystem and syscalls are whitelisted per task. Any dangerous action (file deletion, outbound calls, privilege escalation) trips a circuit breaker and replays the full execution trace. Ships as a CLI plus a local daemon that plugs into existing agent workflows.
Help cross-border freelancers compare stablecoin payout routes, fees and compliance scripts
Problem
Cross-border freelancers wait 3–5 days for bank transfers, lose a few percent to fees and FX conversion, and sometimes get stuck in compliance review. Payment platforms take their own cut and can freeze balances, so more freelancers ask for USDT/USDC — but they don't know which chain to pick, how to estimate net proceeds, or how to explain it to clients.
Solution
A payout planning tool for freelancers: enter client country, amount and currency to compare actual net proceeds and arrival time across banks, payment platforms and major stablecoin routes. Generate a one-page English payout explanation template for clients, plus risk notes and a freeze-scenario checklist.
An auditable execution gate for local coding agents so they can't silently delete or push
Problem
Agent tool calls — Claude Code running shell commands, writing files, calling MCP tools — mostly execute by default, and developers only find out afterward what happened. Subagent drift, accidental deletions and stray commands are increasingly common, while existing safety tooling is cloud-heavy and enterprise-oriented.
Solution
A locally running execution gate that intercepts agent tool calls, applies configurable rules (path allowlists, command denylists, read-only sensitive dirs) to allow, ask or deny, and keeps a full audit log. Deep integration with popular agent CLIs with zero cloud dependency.
Track quota reset patterns for Codex and other agents and get pinged when the window opens
Problem
Vendors like OpenAI run quota reset mechanics for coding agents (e.g. the 28-day Codex reset challenge) but reset times are irregular and there is no official calendar. Developers don't know how much quota they have left or when the next reset is coming, so they miss windows.
Solution
A lightweight calendar and alert service that backfills historical reset records, shows an on-time percentage chart, and notifies you via email, webhook or browser notification when a reset is near or your quota is almost used up. Manual annotations and community-crowdsourced updates are supported.
Command Claude, Codex and Gemini coding agents from one place and let them share findings
Problem
Developers routinely run several coding agents at once (Claude Code, Codex, Grok, etc.) but none of them knows what the others are doing, so the human manually copies insights between them. Subagents also drift and ignore instructions, causing rework.
Solution
A lightweight local console that auto-discovers and registers agent sessions on your machine, providing a shared task and findings board. You can push one agent's output to another with one click, and set guardrails plus status alerts for subagents to reduce drift.
Build a local semantic index of your repo so coding agents stop burning tokens on code search
Problem
Developers using Claude Code, Codex and similar agents find that agents spend a huge amount of time grepping through code with regex tool calls, burning tokens and adding latency while producing unstable results. The search-first approach assumes an unknown environment, but real work happens inside a known codebase structure where repeated searching is avoidable.
Solution
Generate and maintain a local semantic index of your repository and expose it to agents through a lightweight MCP service. Agents query the index to locate symbols, call chains and files directly instead of running dozens of regex searches. The index updates incrementally on your machine and never uploads code. One-command integration with popular agent CLIs.