DeepBlend
DeepBlend 0.13.0 lets moved or tilted thick-walled cups keep resizing, preserving placement and adapting the default main camera to actual geometry; old versions, manual look/handle settings, photography and animation remain protected.
DEV TRENDS / DAILY DISCOVERIES
Daily discoveries from developer communities, independent makers, and open source.
DeepBlend 0.13.0 lets moved or tilted thick-walled cups keep resizing, preserving placement and adapting the default main camera to actual geometry; old versions, manual look/handle settings, photography and animation remain protected.
Lightweight desktop browser reusing the system-installed WebView2 engine, with GUI and headless builds around 4.5 MB each, plus a headless automation browser drivable by AI tools via MCP. Ships a Windows installer.
AI agent skill marketplace to browse skills by category (web/UI, images, presentations, video, documents), preview each skill's examples and description, then download and install them into existing agents like Claude Code, Codex, or Curso…
Free open-source macOS multi-format file viewer that shows Markdown with charts and formulas, PDFs, images, JSON, CSV and source code in one workspace, with Apple Silicon and Intel builds.
Takes a technical concept and generates isometric mechanical animation MP4 videos and interactive HTML diagrams explaining the principle, with no installation required.
Upload an authorized photo to generate a fictional 10-second vertical MP4 goal highlight with Kling O3, keeping your face, clothes, and shoes. No prompt/reference video; crowd audio included; $9.99 each.
Upload a selfie to generate a 10-second vertical horror video with sound in which a ghost crawls out from behind your shoulder; pick a slow-approach or sudden-jump scare for a $4.99 one-time fee, no filming or install required.
An AI video generation site by studio PlusFrame: turns text, images or reference assets (image/video/audio) into video with native audio, using Seedance 2.5 or 2.0 tiers (Mini/Fast/Standard), 5-30 seconds up to 1080p. The "Seedance 3.
Upload two photos to place you and a friend at the keyboard and laptop in a white studio, generating a 10-second duet video in 16:9 or 9:16 with no prompt writing, $9.99 per video.
AI video/image generation studio: Kling 3.0, Veo 3.1, Seedance 2.0, Wan 3.0, Hailuo 2.3, GPT Image 2, Seedream 5, Grok Imagine; Motion Control maps sub-30s dance clips to any character, plus 10s one-photo dance templates.
Upload a photo of yourself, your pet, or your child to make a 10-second vertical MP4 of the 'Bad Times' argument meme with its original audio. Pets stand and point at the camera; people use baby-style pointing and shrugs. $9.99 each.
Remaker AI is an AI image generation, editing and upscaling workbench supporting models such as Nano Banana 2/2.1/Pro, GPT Images 2.5, GPT Image 2, Seedream 5.0, Flux Kontext, Qwen Image 2.
Jean Phil AI makes 10-sec vertical videos with sound starring a fictional blond, curly-mustached gentleman, mixing 5 scenes (Paris flat, dance battle, gallery, café, haunted house) and 4 actions, no photo upload, $9.99 each.
Upload a photo to place yourself in a street performance spot outside a red-light restaurant, generating a 10-second vertical MP4 with sound, from $9.99; not affiliated with Bulin 47, Kevin Wai, or their labels.
Upload an authorized photo and Kling O3 regenerates a 10s vertical MP4 street freestyle scene with people, street setting and new ambient audio. No source video or original track, fixed scene, no custom prompts, $9.99 each.
Upload your photo and up to two friends' to make a 10s vertical party video: you center stage under a chandelier toasting, friends cheering behind, original reference track. 40 credits ($9.99); no Drake affiliation.
开源 Photoshop 替代品 PhotoCraft 的网页版入口与桌面版下载站,纯 Rust 重写,支持图层、蒙版、调整图层、图层样式、文字、矢量、画笔与真实 PSD 文件,编辑在浏览器本地完成,桌面版覆盖 Windows / macOS / Linux;站点另加付费 AI 出图与批量修图(缩放、格式转换、压缩、加水印、重命名,每批前 10 张免费),非官方站点
OriginalTwo-person photo-to-video tool that turns two portrait photos into an orange-studio performance video, with left/right placement, optional action prompts, quality options, and MP4 preview and download;
AI image and video platform for creators: text-to-image and image-to-image generation, popular video effects like Michael Jackson, Raindance and Hotel Lobby, plus watermark removal, quality enhancement and filter restoration.
Free screenshot text editor that uses local OCR to detect text and erase or replace it, running entirely in the browser with no uploads or watermarks, suited to cleaning up screenshots, posters and product images.
Create a personalized song featuring the recipient's name in minutes for birthdays, anniversaries or Christmas, then send it through a dedicated page.
Upload photos spanning from infancy to now to generate an AI growth video with story-driven transitions and your own music, up to 2 minutes long.
Free, no-signup English get-well message picker: filter editable sample wordings by relationship, occasion, tone and format, with teacher card PDFs and workplace sympathy phrasing practice.
English folklore-entertainment web app converting Gregorian birth and conception dates into lunar age and month to query a traditional gender chart, plus a lunar-age calculator and free party-game PDFs. No signup; in-browser; not medical.
GemiAi is a multi-platform AI chat and image workspace: it writes, understands documents and images, generates or edits images from descriptions, and builds web pages.
AI tool that turns two solo photos or one duo photo (people or pets) into a 21-second 720p 9:16 vertical two-person dance video with no prompt writing; login and credits required.
Multi-model AI chat with a personal knowledge graph: switch models within one conversation while keeping context, hear replies in multiple voices, and get automatic summaries, tags and topic links after each chat for search and graph brows…
Indie Radar: low-noise feed for indie/overseas makers. Dedupes first-hand creator sources with local pgvector, scores them on 5 axes, daily curates 2-4 high-value tools and monetization cases, sent via Telegram and GitHub Markdown.
Chrome extension that saves WeChat public account articles as local Markdown, processing the text in-browser without uploading it; exports import into Obsidian, Siyuan, or Typora.
A design-focused iPhone life ledger for long-term records. It learns your spending habits from local entries, suggests categories and next purchases, and revisits life through a bookkeeping map replay and monthly reviews.
A native macOS monitor for Codex usage: check remaining quota, reset time, session tokens, cache reuse and context usage from the menu bar, plus a floating quota badge that follows the Codex desktop pet. Runs locally with no extra API key;
Open-source SaaS boilerplate: one command scaffolds a Next.js + Hono gateway, Better Auth, Creem/WeChat/Alipay payments and generated Postgres RLS multi-tenancy; runs locally on an empty .env, with AGENTS.
A browser game discovery directory curating 100 third-party titles, filterable by control scheme, pace, and player count; four are tagged as AI-assisted with creator attribution.
Browser-local video frame extractor: keyframes, scene change, every N seconds, per-frame, fixed count, FPS, single frame; timeline range select; PNG/JPG/WebP export to ZIP or contact sheet; no upload.
TenYears is a web tool for backtesting dollar-cost averaging into US stocks and ETFs: set custom dates, amounts and intervals, compare up to 4 tickers, use daily closes with splits and dividends, switch dividend reinvestment, and view XIRR…
Free, open-source yt-dlp desktop GUI for downloading video, audio, playlists, channels and subtitles from YouTube and 2000+ sites. Features 4K/HDR/60fps, batch and queued downloads, SponsorBlock, subtitle/cover embedding, 30 UI languages.
A web tool that turns two photos into a 15-second zombie-trend video with sound, following a fixed four-shot sequence. Choose 480p or 720p and 16:9 or 9:16. One-time credit purchase with no subscription;
A skill that generates interactive, well-designed architecture diagrams directly from your codebase, explaining the design in plain language so you get a diagram you can trust. Works with Claude, Codex and other mainstream AI tools.
Skill for Claude Code/Codex: turns Bilibili, Douyin, YouTube or Xiaohongshu videos into Markdown lecture notes with key points, knowledge map, quizzes and video timestamps. Uses subtitles or local faster-whisper; no paid APIs.
11 MIT-licensed English practice sites: IELTS listening synonym drills, reading matching (120 passages), writing correction, 128 AI speaking mock tests, plus illustrated KET/PET/FCE and AWL 570 flashcards. Free, no signup, mobile-friendly.
DAC connects DeepSeek / Pi agents to your product through a REST API, supporting both single agent nodes and agent clusters. Free, MIT-licensed and self-hostable.
Third-party pay-as-you-go API for reading public X (Twitter) data without an X developer account or linking your own account.
DamaYa is a web tool that automatically blurs faces, license plates, and text in photos and videos, offering mosaic, blur, solid masking, eye bars, and stickers.
Free Chinese web tool that turns task, source, time, timezone and failure handling inputs into copyable Grok Bot Routine instructions, with three editable examples and keyboard support.
Online AI transcriber for audio/video files and links, producing timestamped editable text in 100+ languages. Speaker detection, AI summaries, DOCX/TXT/SRT export, text-driven editing to delete or regenerate lines. Free, no signup.
Windows/macOS desktop app that publishes Agent-written Markdown to 41 platforms (Zhihu, CSDN, WeChat, Xiaohongshu, Dev.to). Logins stay local; ships as MCP endpoint and CLI for Claude Code/OpenClaw. Open-source skills plus platform rules.
Open-source TypeScript full-stack generator with 3,241 verified stack combos (React 19, Next.js, TanStack Start, Hono, oRPC, Drizzle, Better Auth). Used via CLI, web configurator, or AI coding-agent skill. End-to-end type safety and AGENTS.
An AI Skill plus companion site for parents to make printable bilingual picture books: age-based content suggestions, single-page cards or stories, page-by-page generation with consistency checks, bilingual scripts and A4/A5 print PDFs.
A CLI that pipes and parses LLM SSE streams (/v1/chat/completions, /v1/responses): it decodes each framed event and prints only the newly added content, one delta per line, plus TTFT, total time and tokens/sec.
macOS desktop app that captures system audio (not the mic) and displays real-time bilingual subtitles: original speech plus Chinese translation.
macOS utility for external displays: uses DDC/CI hardware control for brightness, volume and input source, and falls back to CoreGraphics gamma-table dimming when DDC fails.
Numbered catalog of motion graphics, camera moves, UI animations, web effects and photo styles, with runnable examples in a browsable gallery and an AI skill (Claude Code, Codex) to pick, reuse and port them into projects.
A Chinese-language prompt reference collecting 927 Opus 5.5 video examples. Browse by result, then read the original authors' public prompts to learn how to describe motion effects, product demos, explainer animations and 3D scenes.
HTTP API rendering URLs, HTML or Markdown to PNG/JPEG/WebP, PDF or scroll videos via headless Chromium, with lazy-load scrolling, cookie/ad blocking, device presets, batch jobs and signed webhooks; includes an MCP server for AI agents.
An AI prototyping workbench that reuses old with new HTML. Installed as a Skill, it needs an AI assistant able to read/write files and run commands, plus Node.js 20+; no npm dependencies at runtime, MIT-licensed.
Open-source AI workflow desktop app for macOS, Linux and Windows. Describe a need in natural language; it builds a workflow of functions, LLM agents, schedules and approvals. State persists in local SQLite. No API key;
A Skill that lets AI capture a Perfetto trace from an Android device and analyze a given process's CPU usage, returning top processes, threads and slices plus hotspot and optimization hints, replacing manual work in ui.perfetto.dev.
Uses AI to reverse engineer an original application's binaries and reimplement an AI-rewritten version of it.
Google's official detection tool: upload a file and it checks whether it contains a SynthID watermark, indicating the content was AI-generated.
A TypeScript compiler written in Rust and implemented with AI assistance, reportedly 100% compatible with the official TypeScript compiler.
Browser-based graphic design tool aiming to be an open-source Canva alternative; currently ships around 10 built-in templates.
HeadOrbit uses the AirPods' built-in motion sensors to detect head and neck movements, so nodding, shaking or tilting your head can control a Mac without any camera.
qrypt is a cross-platform CLI tool that mounts cloud drives as local folders, supporting major Chinese cloud storage services so files can be accessed like a local disk.
A Cloudflare-based VPS and virtual host management platform that lets you log in over SSH and administer your servers directly from the web.
A Rust charting library that produces SVG and image files directly without a browser or Node.js, with a charting syntax inspired by ECharts.
Open-source drawing board for macOS, suited for annotating screenshots.
Caplo is an open-source macOS screen recorder that automatically zooms in on clicks, with the view panning to follow the cursor.
Open-source image super-resolution tool that upscales images up to 4x with selectable AI models, processing everything locally in the browser.
A website that checks your browser fingerprint, shows how much personal information is exposed, and points out the problematic areas.
Open-source Android app: install it, pair with your computer by scanning a QR code, then write scripts on the computer to control the phone, with no USB debugging or root required.
A ~$7 ESP32-S3 board plugs into an iPhone's USB port, letting AI models control the phone for automation, with no Mac, developer account, or jailbreak needed.
Open-source macOS app that keeps your Mac awake while AI coding agents like Claude Code and Codex run, with lid-closed wake, overheat pause and low-battery exit.
Open-source, self-hosted A-share data lake that ingests 55 datasets locally—market quotes, financial statements, fund flows, corporate events, macro, futures and options—queryable via Python, SQL, a web console, and MCP.
A community project uses AI to rewrite the 1995 single-player game Xianjian Qixia Zhuan (Chinese Paladin) from scratch in TypeScript, playable in the browser with the main story running through to the ending.
A repository with a step-by-step guide to enabling the fcitx5 + Rime Chinese input method on Debian, covering the operations for each configuration step.
APPLORE is a collection of over 18,000 app icons that supports browsing by category and searching.
A website for looking up the origins of Chinese characters, showing glyph evolution for common characters: oracle bone script (1,425), seal script (10,249), traditional and simplified forms.
Open-source Codex Skill turning a brand logo into a set of paper stickers. Includes three real examples (15 transparent PNGs) plus prompts; keeps logo colors and shapes while giving characters small tasks. Needs an image tool; MIT.
Open-source reading tool that zooms papers, web pages and files into five levels, from overview to a paragraph, and folds Claude Code and Pi replies. PDF side-by-side view, annotations, interactive HTML export. macOS/Windows, pip install.
GitStats is a pure-Python CLI that turns any Git repo's full commit history into a static HTML report: activity heatmap, authors, code ownership, hotspot files, tags, LOC trends, and AI-assisted commit counts. Runs locally.
Uses macOS eslogger to log what AI coding assistants like Claude Code and Codex do on your Mac: commands, files, secrets and network, flagging .env, SSH keys, keychain and risky commands. CLI plus menu bar app, local Markdown reports.
Open-source local-first ebook and PDF reader built with Vue 3, TypeScript and Tauri 2. Annotations, offline listening, progress; papers get a separate workspace with AI Q&A and human edits stored apart. Optional AI, OPDS/Calibre, WebDAV.
Open-source desktop coding agent for Claude, ChatGPT and Grok subscriptions: matches Codex CLI on Terminal-Bench 2.1 with 63% fewer tokens. Main model plans, cheap models work in parallel;
SoftJL Community is a self-hosted AI workbench for software engineering, UML, coursework. It links requirements to models, design, tests, and DOCX specs, traces change impact and coverage, and offers read-only MCP to external coding agents.
Free, no-signup IELTS listening site: 2,069 audio multiple-choice drills on paraphrase keywords from all 40 Cambridge IELTS 11–20 tests, 3,654 British-accent recordings, words hidden by default and local progress. Pure HTML/JS + Vite.
Free open-source macOS app mirroring Android emulators, Android devices via USB/Wi-Fi and iOS simulators in one window, with touch input, control panel, live logs, screenshots, recording, drag-drop installs and 3D iPhone Duo view.
Chromium MV3 extension that only runs on Claude domains: masks timezone, language, Chinese fonts, WebRTC IP and domestic browser UA to mimic a non-mainland-China environment, and shows before/after risk scores from a 10-item weighted model.
Local-first AI assistant for university students: import a timetable once, then ask for weekly classes, tasks or grade reports. Handles document generation, to-dos and persistent memory; terminal, web, QQ and Obsidian share one local store.
TypeScript stdio MCP server wrapping Doc2X cloud parsing of PDFs and scanned images into MCP tools with Markdown, LaTeX, DOCX and layout JSON export. Ships a Skill for submit/wait/download; needs Node.js 18+ and a paid API key.
No project description yet.
Cursor plugin specification and official plugins
The job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
A JavaScript userscript that retrieves direct file download links from eight Chinese cloud drives (Baidu, Aliyun, China Mobile, Tianyi, Xunlei, Quark, UC, 123), modified from the Netdisk Direct Link Download Assistant.
OI Wiki is a community-maintained knowledge base for competitive programming (OI/ICPC), covering algorithms, data structures, math, and contest problem-solving for learners and contestants.
AI Agent Assistant & development framework that integrates lots of IM platforms, LLMs, plugins and AI feature, and can be your openclaw alternative. ✨
A plugin-based, customizable, ad-free music player offered for free.
📄 Awesome OCR multiple programing languages toolkits based on ONNX Runtime, OpenVINO, MNN, PaddlePaddle, TensorRT and PyTorch.
Self-hosted async REST API, MCP server, CLI and web console for TikTok/Douyin posts, profiles, comments and playlists plus watermark-free video downloads, with a self-healing identity pool and PostgreSQL archive;
A curated list of Chinese independent blogs for browsing and discovering independent Chinese-language blog content.
RuoYi-Vue3 is the official RuoYi repository: a front-end/back-end separated permission management system built on Spring Boot, Spring Security, JWT, Vue3, Vite and Element Plus.
MiniOB is a compact database that assists developers in understanding the fundamental workings of a database.
An accessible, humorous Java learning guide covering Java basics, concurrency, the JVM, enterprise development, and Java interview essentials.
Provides multiple Shadowrocket rules with strong ad-blocking capabilities, rebuilt daily at 8 AM.
Hello Algo is an animated, illustrated tutorial on data structures and algorithms with runnable code, available in Simplified Chinese, Traditional Chinese, English and Japanese, with implementations in Python, Java, C++, C, C#, JS, Go, Swi…
Palomino is a fast, GPU-rendered agent workspace. Run agents side by side with your terminals, editor, and browser in one window.
You know when you’re out and remember the one thing that’s sitting on your Mac at home? That’s why we made Mermaid. You just text your Mac from your iPhone, like you’d text a friend, and it actually does the thing.
Helicon gives Muse Code, Meta's coding agent CLI, a proper home: projects and sessions in one sidebar, readable approvals, inline diffs and your plan meter. Free, MIT, no API key. Windows (no WSL), macOS, Linux, VS Code.
Speech-to-text built for conversations with more than one voice with highest accuracy than any in the industry. Lowest diarization error on several benchmarks. 45 languages, up to 25 speakers.
Tidy is an open source alternative to Figma and Paper for designers, engineers, and AI agents. Design interfaces directly from your coding harness (Claude Code, Codex, Opencode, etc), explore the source code, and adapt the tool to fit how…
GBrain Tools is a password manager for AI. Connect Gmail, GitHub, Stripe, Notion and many more once, then use them from ChatGPT, Claude, Claude Code, Cursor or your terminal. No AI holds a password, and every action lands on one log.
Toad lets Claude, Claude Code, Cursor, or any MCP client run commands and work with files on all your machines through one MCP server. Install a small service on each machine. No open ports, and every change asks for your approval.
Put a QR code on your site, receipt or status page. People scan it with the free Nudgebox iPhone app and choose what to hear about. No account, email or phone number. Your backend reaches them with one HTTPS call. No SDK, no device tokens.
iKnowMyMac is a native Mac app that tells you why your Mac is slow, what's filling your disk and what's draining your battery, with the evidence and one action for each. It also updates your other apps from one tab.
Jockey is the first AI that understands your entire media library across photos and video. It analyzes thousands of hours of footage so you can find and connect people, moments and context across videos.
Send webhooks reliably from your infrastructure. Retries, cryptographic signing, health tracking, payload transforms, recipes for Slack/Discord/PagerDuty, and a self-service portal — in one Go binary with PostgreSQL as the only dependency.
Skisso is a free Chrome extension for designers and developers who build with AI coding agents. Point at any element on your live product and edit it in your own design system: your tokens, your components, a shadcn/ui kit that matches you…
Atelier makes your Claude plan last: Codex finds the problems and Claude only judges them (about 71% fewer Claude tokens per review), long chats hand over to fresh sessions, and every session starts from a short brief.
Give Neophyte any product URL and watch an AI agent onboard from scratch. It creates an identity, signs up, verifies its email, navigates setup, and attempts activation.
Scalebrowser is self-hosted browser infrastructure for AI agents. It gives each agent a persistent, isolated identity, keeps logins and passkeys intact, handles captchas and verification, and runs on hardware you control.
I've refreshed StarSkirmish Bench, a benchmark where LLMs play StarCraft BW by writing code, with results from the flagship models of OpenAI, Anthropic, Google, and Muse
Sygal converts PDF, Word, Excel, images and web pages into clean Markdown and JSON, locally on your Mac or PC. One-time purchase, no subscription.
Research communities for independent AI agents, with problem-specific discussion and reviewed papers.
Deploy open-source apps on a dedicated server. Pick an app, a region, and a plan. Pay by the hour, capped monthly.
Structure-first operating process for AI coding agents: Blast / Mid / Inner build altitudes with clear ownership, then T1 structural, T2 bug/ops, T3 adversarial quality gates.
I build websites for a living. Lately, it's more about prompting than actually writing the code (the art part is still there). So I decided to leave that part to Claude too and see what it would come up with.
grimoire.bio — a space for your biology questions.
We noticed during the course of many many years of DBA related work that fixes to database issues and queries are always reactive. An entire database can fall over extremely quickly when an unbounded query grows without warning.
Do you remember disc? Then come check out the new IndieStation!
Endo is a Higher-Order Harness (HOH). From the easiest concept to understand to the more complex ones, Endo gives you: * A single, simple UI to use any agent provider with your existing subscriptions, and hot-swap agents right in the middl…
Issue tracking isn't dead. It's moving closer to code. Gitsu is a fast, AI-native, keyboard-first Mac and web app on top of your existing GitHub Projects and issues. Everything syncs with GitHub in the background.
Long story short it's a media library licensed under Apache2. It cares about your privacy. It doesn't send any data and it doesn't have telemetry. It is local only project. It'
I built this for myself. I often have questions I want quick answers to such as "how cold did it get last night?", "I just put down fertilizer, when will it rain next?", "is today unusually cold or warm for this time of year?
Discover classic books through their opening words. Read, reveal, and save your next favorite.
No project description yet.
Browser game where you coach the final two minutes of real NFL games: pick run or pass, set routes, targets and defense. Each call is graded by estimated win probability, not just the result. Daily and Survivor modes; nflverse data.
Python CLI that turns audio files into TXT, SRT, VTT, TSV or JSON text and subtitle files using local Whisper models or the OpenAI speech API, with optional DeepL translation. Requires Python 3.11+ and FFmpeg; MIT licensed.
A Pokemon-themed tower defense game.
Fan-made Roadster side-scroller web game built with Three.js on Cloudflare: fly over ground traffic, land to avoid airborne cars. Switch with Space/E or touch buttons;
A mini program that offers one sentence each day; users can draw an image for the sentence and export it to share.
Unofficial NanaZip fork tracking upstream: live theme switching, per-archive code-page switching for legacy filenames, IEC size units, UI language option, Ctrl+W close. Package Kaesinol.NanaZipRev installs alongside official build.
A free decision-model API, Bao Qingtian: trains in under 1ms on integrated graphics, scores 38.10% on the jev benchmark, uses a fully random algorithm whose fallback keeps choice probabilities summing to 1. Two API keys are open.
Vite + React + TypeScript + Tailwind CSS v4
Turn two photos into a 10–15s zombie-couple reunion video: pick survivor and zombie, and the pair recognize, embrace, then flash back. Credits required, packs from $6.99, no free trial. Uses MiniMax H3 and Seedance 2.0 Mini.
A discussion on whether developing projects on GitHub and using Claude Code through an external codespace would prevent account bans; the poster says Codex underperforms while Claude Code has already gotten several of their accounts banned.
Try another keyword or category.
19 repositories were already featured in the last 3 reports, so they are not repeated in today’s discoveries.
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
XERJ is the new way for AI to search data. Its autoindex capability activates agents to know your data without the token waste of grep and sed.
Official, Anthropic-managed directory of high quality Claude Code Plugins.
DEV TRENDS / COMMUNITY
Discuss this edition
Sign in with GitHub to share a thought, question, or useful follow-up.