DEV TRENDS / DAILY DISCOVERIES
Daily discoveries from developer communities, independent makers, and open source.
StoryX is an interactive content creation platform that turns a creator's images, text and videos into an interactive world fans can enter, experience and engage with, bringing the content to life.
Browser 3D tool for custom dirt bike graphics: paint each plastic, rotate to preview, pick make/model/year, add number plates, upload logos/photos or AI-generate full sets; export per-plastic PNGs and a full-set PDF. Free design.
Use AI to seamlessly merge two or more photos into a natural scene. Easily add people to group photos, change backgrounds, or blend product images online for free.
Long-term market watching dashboard with a watchlist overview, weekly and monthly charts, and custom indicators, built with Next.js, KLineChart, and SQLite.
A desktop ambient companion: gently looping animated illustrations with one-to-three-line calming phrases that auto-switch by time of day (morning, noon, evening, late night, weekends).
Enter any Chinese character to see its evolution from oracle bone and seal script to traditional and simplified forms, with Middle Chinese fanqie, Old Chinese reconstruction, Shuowen Jiezi glosses, and phonetic-series families.
Unofficial archive of Justin Sun: a timeline of key experiences and public events, 146 audio episodes of The Road to Financial Freedom Revolution with full transcripts, plus online reading of This World Is Both Cruel and Gentle.
AI multi-role dubbing workbench: paste a script to auto-detect roles and assign voices, adjust emotion per line, export finished audio. 9 languages incl. Chinese, English, Japanese, Korean; free open beta, no sign-up.
AI scientific-diagram tool that turns one sentence into labeled, editable illustrations for papers and slides, with multi-discipline templates and PNG export; free during beta.
AI game asset studio for solo and indie game developers: generates and manages style-consistent sprites, animations, sprite sheets and matching icon sets for one ongoing game project.
AI image generation and reference-image editing website for product scenes, fictional character storyboards and posters with text;
AI platform for creating images, video and music, with multi-model generation and image editing; after login you get 20 free FLUX.1 Schnell images daily, other models are credit-based.
AI photo-to-video tool: preview and pick a scene template, then generate a 15-second video with original AI music. Pay per video with no subscription.
Library of 18,000+ real App Store and Google Play icons plus each redesign's history, browsable by color, style, category. Free no-signup icon export, screenshot generator, PNG-to-SVG; AI icons via credits.
Browser-based interactive 3D solar system: orbit the Sun and the eight planets, toggle between real distances and true scale, and land on the Moon and Mars. Free, ad-free and no sign-up; suited to classroom demos.
LogoMovo is a browser-based logo design workbench for exploring shapes, fonts, colors, and typography, then exporting designs as SVG or PNG;
AI nail design planner: generates downloadable reference images by color, length and detail, and previews styles on hand photos. AI requires login; free plan has 5 AI actions/month, plus paid tiers; previews may differ from salon results.
Browser photo editor for restaurant menus and delivery listings: tune light, color and crop, compare with originals, download JPEG. Manual edits need no account; AI edits need sign-in and credits. Check ingredients, portions, plating.
Upload 1–3 product photos, pick a preset, color, or scene description, review each, and export square or vertical PNGs. Free trial via invite code; public samples need no signup. AI may alter product details.
Personal AI photo styling preview: adults try hairstyles, outfits and portrait scenes on their own photos, compare and download results. Free preview requires login, invite code and credits; samples are AI-generated.
Browser-based platform of utilities, mini-games and original comics: image compression, Markdown-to-HTML, EPUB reading. Basic compression and some samples are free; full tools or content require login and credits.
AI two-person dance video generator: upload two frontal portraits and both dance a fixed 1980s synthwave routine in sync, output as a 14s 9:16 vertical video, no prompts needed. Credit-based one-time pricing, 480P–1080P, from $4.99.
One-click document translation that keeps the original layout. Upload a PDF or Word file, translate into 20+ languages, preview the first 3 pages free, and pay only if satisfied.
Live Photo maker and converter: turn videos or GIFs into Live Photos on iPhone; the web app previews clips, exports GIFs and video stills for free, plus credit-based AI photo animation.
AI-tone detector for web fiction: paste text for a 0-100 score, flagged sentences, and a ready-to-use de-AI rewrite prompt. Runs fully in the browser, no uploads, free, no signup; plus a 38KB single-file offline version.
Quark/Baidu Netdisk organizer + Agent Skill: paste a share link to auto-detect the code, save to your drive, archive by year/month/day, and create new links. Local captions need no LLM API. Docker-deployed web UI with SKILL.
An AI API relay that lets you call OpenAI, Claude, Gemini, Grok and other models with a single key.
Unofficial open-source browser fan remake of Scam With Your Friends, unaffiliated with its developers. Play online with up to 3 friends in a 3D office, taking AI callers and hitting daily quotas. React 19 + Three.
Browser extension that sniffs and batch-downloads video, audio, images, subtitles and docs from Bilibili, Douyin, Xiaohongshu, X and more, with preview, anti-hotlink proxy, MSR/MSE recording, Blob capture and cloud-accelerated downloads.
MIT-licensed browser extension that captures a chosen tab's audio to generate AI-translated speech and synced original/translated subtitles, no existing captions needed. Separate controls for audio, dubbing and subtitles;
Self-hosted Apache-2.0 wiki for engineering teams: docs embed live charts querying Prometheus, PostgreSQL, MySQL or MariaDB. Draft/diff review approval, co-editing, comments, search, optional MCP server for Claude Code/Cursor.
MiroTalk RND is a self-hostable WebRTC random video chat app. Users join a waiting queue and are automatically paired one-to-one with another user; clicking Next immediately matches a new chat partner.
Paste a GitHub URL to get a 0-100 launch-readiness score plus the top fixes to make first. A companion AI skill writes promo copy following the rules of 16 channels, including Hacker News, V2EX, HelloGitHub and Xiaohongshu.
Free web tool to download public videos from X, Facebook, TikTok, Instagram, Threads and Reddit, convert clips to GIF with trimming and compression, and upscale 2x via Anime4K. Local processing, 21 UI languages, plus a Chrome extension.
Protocol set turning 14 engineering roles into a multi-agent delivery pipeline: 3 gates per handoff (file exists, EXITCODE proof, redaction), failure loopback by issue type, 3-layer memory caps. Zero deps, 3 runnable demos, self-iterating.
Local-first desktop app that aggregates AI coding assistant usage from 12 clients (Codex, Claude Code, OpenCode, DSH). Read-only capture into local SQLite, no telemetry; token trends, model rankings and honest measured/estimated labels.
Local AI photo library that writes a readable caption and 1–10 quality score per photo, then daily picks one for a literary title and short essay. Next.js 15, Hono, SQLite, BullMQ; OpenAI-compatible vision models (llama.cpp/Ollama).
Generates anime characters, portraits and scenes from text prompts, with reference images for detail. Styles include cinematic cel, 90s retro, chibi, soft romance and monochrome manga—useful for concept art, avatars and storyboards.
A native, lightweight macOS music player built with SwiftUI that brings Dynamic Island-style controls to the screen notch.
Local-first desktop AI coding agent (Tauri 2, WIP). It reads, edits, runs and verifies work, showing tool calls, diffs and failures; writes need approval and edits are snapshot-rollback. BYO tool-calling model, local encrypted keys.
Browser-based AI audio toolbox in 6 languages: vocal/accompaniment separation, denoising, multi-stem splitting, plus free local video-to-audio, trimming, merging and waveform editing tools. AI features need login.
BossAlign (Roundworm Engine) is a multi-reviewer preference alignment framework for LLM revisions, addressing overcorrection and one-size-fits-all rewrites.
Hindsight is an open-source, local-first computer activity tracker for macOS and Windows. It logs the apps and windows you use in the background and summarizes time spent by day, week, and month for personal review.
iOS app for people with huge photo libraries they never revisit. It randomly pulls 12 photos from chosen albums, years or months; gestures decide keep or delete. Photos can be added to albums and viewed with same-date/location shots.
A free 3D solar system simulator that runs directly in the browser with no download or signup. Orbit and zoom around the Sun and eight planets, change time speed to compare orbital periods, toggle real distances and real scale, and land on…
Generates visual variations of an existing short video ad, helping marketers quickly produce multiple ad creative versions.
Free online tool that turns an uploaded video file into a shareable link, with no signup, no watermark and no re-encoding, so original quality is preserved. Generated links can be embedded in Markdown, HTML, Tencent Docs and Feishu Docs.
An evaluation of JEV on high-concurrency structured decision tasks, covering 9 medical scenarios and 26 tasks, benchmarked against several other models.
APPLORE archives 18,000+ real App Store and Google Play app icons, browsable by color, style and category, saving each redesign version.
China's first systematic FDE directory: 5,600+ practitioners, 529 FDE service firms, 300+ case studies and policy docs, all publicly sourced, updated daily. Offers llms.txt and site-wide Markdown views for AI Q&A.
A web tool for writing personal digital letters from real shared memories: write your own text or edit an optional AI draft, add photos, and preview the envelope-opening view before deciding to share. Writing and preview are free;
Personal essay by someone born in 1999 who once preferred claiming a millennial birth year, but now proudly identifies with the '90s, arguing that decade was better than the 2000s or 2010s. From the 札迹 short-essay series.
Telegram group bot @GroupModAppBot configured via a Mini App: join verification, 7 anti-spam checks with presets, keyword replies, message templates, cron tasks, giveaways and moderation; 11 languages, no commands or self-hosting.
Pinyin input that shows regular characters in seal-script fonts, so text stays copyable and searchable. 13,285 Shuowen Jiezi glyphs, radical and 540-section lookup, Rime schemes for Windows/Android, plus three offline single-file web tools.
Project URL Category JS Project Title See what your local LLM is really doing, on Ollama Project Local observability UI for Ollama.
@/var/folders/bj/m5pmbwnn7r97zp4g7p5bbr40000gn/T/tmp.488aXW0JMF
Local-first workbench that auto-discovers 14 installed AI CLIs such as Claude Code and Codex in one chat-style UI, where agents collaborate via @mention, A2A calls or group chat, with self-hosted Relay, LAN P2P and E2E encryption.
memnest stores coding-AI chats and decisions locally in SQLite, so new sessions can search and read originals to recall past rationale and error fixes. Local BM25 + multilingual vector search;
Free MIT tool offering 48 inspectable reasoning methods with steps, conditions and examples for comparing options and checking assumptions. v1.5.1 instructions in Simplified Chinese, Japanese, English and Korean;
Enter a domain to list all of a site's icons (favicon.ico, PNG, SVG, apple-touch-icon), then download the highest resolution or batch-package them for bookmark navs, link cards and competitor tables.
Unlikely maps cross-domain connections into explorable idea maps for writers, designers and indie developers: see the network, why nodes connect, reselect and reorder, then export Markdown/JSON/SVG. Vanilla JavaScript, MIT, runs locally.
zeroshot orchestrates coding agents via configurable execution graphs: decompose, implement, review independently, and return failed work for rework. Topology sets parallelism and feedback paths; gates prevent self-approval. MIT-licensed.
Vue 3 admin template on TypeScript and Naive UI: 55 business components plus 8 standalone @robot-admin packages for forms, tables, themes and permissions, maintained in SPA, Monorepo, MicroApp and Module Federation architectures.
Samples 5,817 McDonald's stores across 337 Chinese cities via the official McDonald's MCP to compute a 0–100 City McDonald's Index. Includes a pure-function engine, offline snapshot and a single-file offline page.
Cangjie's Tauri-like desktop framework: a statically compiled Cangjie backend with HTML/JS in the system WebView, joined by an IPC bridge and capability allowlists.
Native GNOME (GTK4/Libadwaita) app that imports Markdown, PDF, HTML and EPUB into a linked, cited Markdown wiki, answers questions with numbered references, and exposes the wiki via MCP to Claude Code, Codex and Cursor.
Local-first desktop workspace to run multiple coding agents (Pi, Claude Code, Codex, OpenCode) in one project, managing sessions, agent replies, tool activity, execution status, per-turn diffs, files, terminal and browser.
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
The open agent skills tool - npx skills
Cross-platform Xcode replacement. Build and deploy iOS apps with SwiftPM on Linux, Windows, macOS.
RuoYi-Vue Pro admin scaffold on Spring Boot, MyBatis Plus, Vue and Element, plus a WeChat mini program.
Downloader and data-collection tool for Douyin/TikTok, used to fetch and bulk-collect videos and related platform data.
Production-grade platform for building agentic IM bots on Discord, Slack, LINE, Telegram, WeChat, Feishu, DingTalk, QQ and Matrix, with agent and knowledge-base orchestration plus a plugin system and many model/service integrations.
Open-source, self-hosted alerting for metrics, logs, and databases. Includes a built-in AI agent and MCP server.
Curates and shares interesting, beginner-friendly open source projects on GitHub to help developers discover and learn from them.
Timed task management platform supporting Python3, JavaScript, Shell, and TypeScript.
A text editor for Windows, Linux, and macOS, positioned as a China-made editor.
fscan is an intranet comprehensive scanning tool that enables one-click automated, all-round vulnerability scanning to help security teams quickly discover internal assets and weaknesses.
Searchable encyclopedia of Linux commands, bundling a full command manual with detailed explanations and learning notes for quick lookup and reference.
Automation tool for Bilibili daily tasks such as sign-in, watching and coin tasks. Deploys via Docker, Qinglong panel, K8s and more, and now includes AI features.
Most teams ship, then hope it lands. Articos MCP lets you check first. Right inside Claude or ChatGPT, run your headlines, positioning, or pages past simulated buyers based on your actual customers, and see how they respond before it goes…
A free, open-source Mac app for talking-head videos. A teleprompter that follows your voice, one-button recording of camera and screen, and every take in plain folders.
Most social APIs let an agent publish text or media it already has. PostNitro lets it design the post too. Connect from Claude or ChatGPT in one click (OAuth, no API key), or use our 35-tool MCP server, CLI or REST API.
Kai is an AI executive assistant for solopreneurs, freelancers and startup founders. It works across your email, calendar and tasks: it triages your inbox and drafts replies, schedules around your availability, records and transcribes your…
100% open source, with local models that run on your iPhone or iPad. Transcribe on-device for free, without usage limits, an account or a subscription.
We took the top 3 Product Hunt launches and asked ChatGPT, Gemini and Claude the exact questions their buyers ask: “Best browser for Claude Code?” “How do I track agent run costs?” “Best open-source Grok alternative?” Nine answers.
You started a business, not an accounting department. Fina Business gives founders the essentials of business accounting without the complexity of traditional tools.
AdHelm is a free, Apache-2.0 command-line tool that lets your AI agent connect to your own ad and analytics accounts, report performance, audit them, and recommend changes, entirely on your machine.
Mi Speaks reads your books, articles and PDFs aloud. 2.0 adds 28 neural Studio Voices, American and British, generated entirely on your iPhone: nothing is uploaded, and they work in airplane mode.
Xiryn is a free, open-source personal agent that does what Muse, Grok Bot and Dots do — on your own computer.
A free, open-source macOS app where your own Claude Code, Codex, Copilot or Grok builds your video as a real Remotion project. Watch it change live, click any element to tune it, give notes in chat, export MP4 up to 4K. No timeline.
Vincentt is the fastest way to build and launch camera experiences on the web. Build with your own coding agent (Claude Code, or Codex) and our open SDK, preview live on your phone, then publish to a link anyone can open.
HarnessHive is the AI-native design ecosystem for wiring harness engineers. Schematic, layout, BOM and a to/scale formboard, all in one live document. With an AI assistant that helps you at every step.
A free, open-source diagram canvas: MIT, local-first, no account, in your browser. One canvas, five ways in. Draw it. Write it as code, or paste Mermaid, D2 or Structurizr. Ask any of 10 AI providers, or a local model.
One console for every server you run: web terminal, file editor, scheduled jobs, fleet-wide commands, 150+ install recipes, plus monitoring, phone alerts and status pages. No open port, and every command is signed in your browser.
Connect Codex or Claude, add a speech key, and send your first document to PDF TO MP3.
SolidCraft: an open-source parametric 3D CAD application built in pure Rust. Desktop, web, and automation through MCP.
An experimental pattern-first DAW
Brickbrag is a whole world made of bricks, shared live with everyone online: a town to hang out in, a race track, an airfield, robot-driven buses out to the volcano – and land of your own to build on.
I think a chasm has been crossed. Opus 5.5 few shotted a fully working Figma for me. In the past year I have been trying to benchmark vibe coded apps and see how good they get and this week it did it with very very little guidance from me.
Opensidian is an open-source, vault-compatible markdown notes app for Linux. Wikilinks, backlinks, graph view and Obsidian community themes. Rust + Tauri v2, no Electron.
About 11 months ago i revived the amazing but unfortunately unmaintained spotify-tui. Since then I have been working on it almost non stop, finding a lot of problems on the way.
I created MemorySync because I kept running into the problem of AI agents not being able to remember useful things about users between conversations.
SwiftUI-like components, styles as modifiers, and a compiler that catches broken links before visitors do. How this site is made, and why I love working on it.
Dictation that runs on your own computer. Press a key, speak, and the text lands in the app you are using. Free and open source, for macOS, Windows, and Linux.
Multiplayer AI: Bring your team and agents together
Hey HN, im Marc, co-founder of ObsessionDB. We are open sourcing ChKit, our tooling to manage schemas, migrations and backfills for clickhouse. I am a big fan of the type safety drizzle provides for the database.
Long-term memory for AI coding agents (Claude Code, Codex, opencode) as a tree of plain files
A service and open source tool to get open data streams in two commands. No sign up needed for subscribing. Streams to try: - https://hypermass.io/stream/
Try out Artifacts created by Claude users
The author used Gemini 3.8 to build an AI daily digest that automatically compiles AI-related news each day.
makepaper.app is a wallpaper creation app; its landing page was rebuilt overnight with the AI models Astra6.1 and claude5.5 opus plus skills.
LiveLore is a free horror/suspense fiction site where readers vote on how serialized stories continue. Stories are AI-assisted and labeled with review status;
A web-based Photoshop alternative built on the open-source PhotoCraft project, bringing its image-editing features directly to the browser.
Pure front-end GPT Image 2.5 image workbench: bring your own API key (localStorage) with images in IndexedDB. Streaming preview, 2-4 aspect ratios per prompt, sketches, up to 16 reference images, inpainting, searchable local history.
NonsPlayer music player arrives on mobile, starting with Android (iOS delayed). Offers multi-source audio adapters, a WebAssembly plugin system, local music playback, daily-use UX design, and low memory use (~10-20MB in background).
Mu3ic 1.5 is an Android player for Navidrome/Subsonic, Jellyfin and Emby. Adds 10-band EQ, per-output presets, crossfade, split Wi-Fi/mobile quality, listening analytics, scrobble rules, fast sorting of huge libraries and steadier updates.
Free MIT-licensed desktop DB client (Navicat alternative) for MySQL, PostgreSQL, SQL Server and Oracle. Connections and keys stay local; includes connection management, object tree, data browse/edit, DDL, SQL editor, dump .
Three Shennian tutorials: connect a phone to AI on an online computer, index materials by name/location/read rules so they need no re-explaining, and choose local link vs Docker. Free now; third-party AI billed separately.
Upload a selfie; AI makes an 8s 9:16 MP4 of you blinking and smiling as a grey ghost appears behind your shoulder — slowly or lunging at the camera, with sound. Free 5.5s watermarked preview; full 8s HD $4.
Try another keyword or category.
14 repositories were already featured in the last 3 reports, so they are not repeated in today’s discoveries.
Uses AI to reverse engineer an original application's binaries and reimplement an AI-rewritten version of it.
ArtCraft is an intentional crafting engine for artists, designers, and filmmakers
A spy satellite simulator in your browser, except the data is real. Live open source spatial intelligence on a photorealistic 3D globe.
![[Tool nomination] Ad Multiplier: visual variations of an existing short ad](https://img.devtrends.site/images/bc00c228ddbd5dab7e447faece7b15660f49d4dd98df4e49916595e85901c1ec.png)
DEV TRENDS / COMMUNITY
Discuss this edition
Sign in with GitHub to share a thought, question, or useful follow-up.