# Feyard > A curated matrix of 26 free, browser-first tools — AI image, 3D, CAD, audio, video, text, PDF and OCR utilities plus Sudoku. AI processing runs on-device via WebGPU with WASM fallback; classic pipelines run entirely in local TypeScript — no servers, no sign-up, no file uploads. Feyard is an indie-built suite of micro-tools that run entirely in the browser. Each tool does one thing superbly: background removal, image compression, upscaling, photo restoration, watermark and object removal (LaMa inpainting), photo-to-3D depth reconstruction, web page cleaning, speech-to-text, subtitle generation, text-to-speech, vocal removal and denoising, PDF RAG search, PDF summarization, PDF editing (merge, split, compress, sign), OCR, bitmap vectorization, sketch/photo-to-DXF CAD tracing, 3D model optimization, point-cloud downsampling and conversion, font inspection and subsetting, AI text humanizing, diagram generation, video compression, and Sudoku. Powered by Transformers.js, WebGPU, three.js, Depth Anything V2, Tesseract.js, Kokoro-82M, Whisper, Swin2SR, RNNoise, @jsquash WASM codecs, WebCodecs, and a local-first architecture. Available in 8 languages: English, 中文, Deutsch, Français, Español, Italiano, Nederlands, Português (BR). ## Products - [Remove BG](https://feyard.com/products/remove-bg/): AI background removal, 100% on-device via WebGPU. Three model tiers (Fast / Sharp / Sharp · GPU), a manual edge brush that edits the alpha channel, batch ZIP export, no watermark and free for commercial use. - [Compress](https://feyard.com/products/compress/): Free online image compressor. Batch-compress and convert JPEG, PNG, WebP and AVIF with MozJPEG/OxiPNG WASM codecs; EXIF stripped automatically. - [Upscaler](https://feyard.com/products/upscaler/): AI image upscaling & denoising with two on-device engines (Swin2SR + Real-ESRGAN anime). 2×/4× modes, batch queue, PNG/WebP/JPEG export, no uploads. - [Restore](https://feyard.com/products/restore/): AI photo restoration studio — repair scratches, sharpen blur, colorize B/W photos with DDColor, fill objects, ID photos, smart crop, batch queue. Six tools in one; colorization free at full resolution. - [Depth 3D](https://feyard.com/products/depth-3d/): Free 3D photo maker — on-device depth AI (Depth Anything V2) turns a picture into a cinematic parallax motion video (MP4/WebM, 720p free / 1080p Pro) with four camera presets, or an orbitable relief mesh/point cloud exported as GLB/OBJ/STL/PLY/USDZ plus a depth PNG; Pro adds 700px depth maps and a manual depth brush. - [Vectorize](https://feyard.com/products/vectorize/): Bitmap to SVG/DXF converter powered by a custom in-browser engine — median-cut color quantization, crack-following contour tracing and Bézier curve fitting in pure TypeScript. Logos, icons, illustrations to crisp layered vectors with an editable palette. - [Cleaner](https://feyard.com/products/cleaner/): Turn any web page into clean Markdown. Readability + Turndown, 100% on-device. - [Speech-to-Text](https://feyard.com/products/stt/): Whisper speech-to-text. Upload audio or record live, multilingual, on-device. - [Subtitle Studio](https://feyard.com/products/srt/): Turn video or audio into timed SRT/VTT/ASS subtitles with word-level cues and a bilingual translation track — fully in-browser. - [Voice Synth](https://feyard.com/products/tts/): On-device neural text-to-speech via Kokoro-82M. Dozens of voices, zero servers. - [Vocal Studio](https://feyard.com/products/vocal-studio/): Vocal remover and audio denoiser powered by RNNoise (WASM). Split vocals from background music or clean noisy recordings. - [RAG Search](https://feyard.com/products/rag/): Client-side PDF RAG. Extract, embed, and retrieve — all on-device, with cited answers. - [PDF Summary](https://feyard.com/products/pdf-summary/): Summarize & extract from PDFs with on-device LLMs (SmolLM2 1.7B / Qwen2.5 1.5B). Long-document aware. - [OCR](https://feyard.com/products/ocr/): Image-to-text with Tesseract OCR via WebAssembly. 100+ languages including Chinese, Japanese and Korean; word-level bounding boxes; zero uploads. - [PDF Tools](https://feyard.com/products/pdf-tools/): A free offline PDF toolbox — merge, split, rotate, compress, reorder, delete, encrypt, watermark, render to images, and assemble images into a PDF. Browser-native via pdf-lib and pdf.js; nothing is uploaded, every action runs on the device. - [Sudoku](https://feyard.com/products/sudoku/): Classic Sudoku with AI logic hints, a step-by-step solver, and printable puzzle packs. Four difficulties, random generation. - [AI Humanizer](https://feyard.com/products/humanizer/): Rewrite AI-generated text so it reads human. On-device rewriting (Chrome Built-in AI, local Qwen2.5-1.5B/SmolLM2-1.7B, or an offline rules engine), a 6-signal AI-likelihood score, flagged AI-tell phrases, and a word-level diff. - [Diagram Studio](https://feyard.com/products/diagram/): AI flowchart and mindmap generator. Describe what you want in plain language — a small on-device model writes the Mermaid code and renders a clean diagram live. No uploads, no sign-up, export to SVG or PNG. - [Video Compressor](https://feyard.com/products/video-compress/): Free online video compressor. Shrink MP4, MOV, MKV and WebM with native WebCodecs — H.264, VP9 and AV1 re-encoding runs entirely in your browser, starts instantly, and never uploads a byte. - [Eraser](https://feyard.com/products/eraser/): Free AI watermark and object remover. Paint over a logo, date stamp or unwanted object, or click it with SlimSAM Magic Select (Pro); Auto routes each mark to zero-download fast-marching or LaMa inpainting (51M params, ONNX Runtime Web). Free up to 2048 px, Pro 4096 px and same-position batches of 10. 100% on-device. - [CAD Trace](https://feyard.com/products/cad-trace/): Convert part photos, hand sketches and scanned drawings into editable CAD. A custom on-device contour engine fits native LINE, ARC, CIRCLE and POLYLINE entities on layers, with four-point perspective correction and real-world scale calibration. Exports DXF (R12/AC1009) and SVG; preview is free and unlimited, exports are metered (3/month free, Pro unlimited). No image is ever uploaded. - [3D Model Optimizer](https://feyard.com/products/model-optimizer/): Compress GLB and glTF files 100% on-device. Meshopt/Draco geometry compression, quantization, vertex welding, triangle reduction with automatic rig/morph protection, texture resizing and WebP re-encoding, split before/after 3D preview and an asset report. Accepts OBJ, STL and PLY; free batches up to 10 files/100MB, Pro 100/500MB; nothing is ever uploaded. - [Point Cloud](https://feyard.com/products/point-cloud/): Downsample, crop, filter and convert PLY/PCD/XYZ point clouds in your browser — voxel, random and every-Nth decimation, color histograms, farthest-point and outlier removal, crop and axis alignment, GLB export. 100% on-device, no upload. - [Font Forge](https://feyard.com/products/font-forge/): Inspect, preview and extract font glyphs locally — view every glyph, read the name table, export SVG, and subset TTF/OTF/WOFF in your browser. No upload. - [RAW Studio](https://feyard.com/products/raw-studio/): Develop camera RAW files (DNG, CR2, CR3, NEF, ARW, RAF, ORF, PEF, RW2) in your browser — white balance, exposure, tone zones, curves, HSL, clarity, noise reduction, geometry and presets, with JPG/PNG/WebP/16-bit TIFF export. 100% on-device, no upload, no watermark. - [SQLite Editor](https://feyard.com/products/sqlite-editor/): Browse, query and edit SQLite databases in your browser — open a .db file or start fresh, run SQL with autocomplete, inspect schemas, import CSV and export your database. Powered by SQLite compiled to WebAssembly, fully on-device, no upload. ## Comparisons Migration and alternatives pages for tools that have been retired or paywalled. - [remove.bg alternatives](https://feyard.com/alternatives/remove-bg/): remove.bg retires its standalone site on 1 December 2026 — the web tool moves into Canva, the API moves to Leonardo.Ai, and the Photoshop/Figma/GIMP/Sketch plugins plus Zapier/Make automations are discontinued. Side-by-side comparison with remove.bg, Canva, Photoroom and Adobe Express, a pre-shutdown checklist, and an embedded on-device remover. ## Sudoku Tools - [Sudoku Solver](https://feyard.com/products/sudoku/solver/): Enter any grid and get the solution with a step-by-step explanation. Each placement is annotated with its solving technique (Naked Single, Hidden Single). - [Printable Sudoku Packs](https://feyard.com/products/sudoku/printable/): Free printable Sudoku puzzles formatted for print or PDF. 16 curated packs (4 difficulties × 4 packs) plus a random pack generator. Each pack has 6 puzzles and an answer key on a separate page. ## Preview - [VocalFlow](https://feyard.com/products/vocalflow/): Preview landing page for an AI text-to-speech reader that reads any tab aloud in 40 languages with 80+ voices. ## Guides Long-form technical articles on how the on-device pipeline actually works. - [Why AI Tools Should Run in Your Browser](https://feyard.com/blog/why-ai-tools-should-run-in-your-browser/): The case for local-first AI — privacy, cost, latency and offline capability. - [WebGPU vs WASM, Explained](https://feyard.com/blog/webgpu-wasm-on-device-ai-explained/): How on-device inference is scheduled, when WebGPU is used, and what the WASM fallback costs. - [Private Speech-to-Text Guide](https://feyard.com/blog/transcribe-audio-locally-private-speech-to-text-guide/): Running Whisper locally for transcripts that never leave the machine. - [Compress Images Without Uploading](https://feyard.com/blog/compress-images-without-uploading/): WASM codecs, AVIF/WebP targets, and batch compression in the browser. - [Extract Text from Images in Your Browser](https://feyard.com/blog/extract-text-from-images-in-your-browser-ocr-guide/): Tesseract.js in WebAssembly — page-segmentation modes (PSM 3/6/7/11), preprocessing, and auditing accuracy with confidence scores. - [Remove Image Backgrounds in Your Browser](https://feyard.com/blog/remove-image-background-in-your-browser-guide/): MODNet vs BiRefNet Lite, why a fixed 512×512 graph decides output quality, and why RMBG-1.4/2.0 is excluded on licensing grounds. - [Summarize a PDF Without Uploading It](https://feyard.com/blog/summarize-pdf-locally-without-uploading-guide/): pdf.js extraction, 8,000-character chunking, map-reduce summarization, and question mode. ## Use Cases End-to-end workflows that chain several tools into a finished job. - [Use Cases Index](https://feyard.com/use-cases/): All published workflows. - [Paper to a Searchable Archive](https://feyard.com/use-cases/paper-to-searchable-archive/): Digitise a box of documents into locally searchable text without uploading a page. - [Bilingual Subtitles Without Uploading](https://feyard.com/use-cases/bilingual-subtitles-without-uploading/): Transcribe, review with timestamps, translate, and export dual-language SRT. - [Local Product Photo Pipeline](https://feyard.com/use-cases/local-product-photo-pipeline/): Cut out, repair, upscale and compress a catalogue in the right order. - [A Private Reading Stack for Researchers](https://feyard.com/use-cases/private-research-reading-stack/): Triage papers, interrogate the shortlist, and search a local corpus. ## Localized Guides Articles written for a specific market, not translations. - [Français] [Numériser ses documents sans le cloud : le guide pratique RGPD](https://feyard.com/fr/blog/numeriser-ses-documents-sans-le-cloud-rgpd/): Document digitisation under the GDPR — sous-traitance, transfers, retention periods under the Code de commerce. ## Company - [About](https://feyard.com/about/): The local-first mission and principles behind Feyard. - [Blog](https://feyard.com/blog/): Deep dives on browser-based AI — private STT, local image compression, WebGPU/WASM internals and the local-first movement. - [Contact](https://feyard.com/contact/): Reach the Feyard team. - [Privacy Policy](https://feyard.com/privacy/): How Feyard handles data — short, readable, no legalese. - [Terms of Service](https://feyard.com/terms/): Terms of use for Feyard tools. ## Optional - [Sitemap](https://feyard.com/sitemap.xml): XML sitemap of all pages across 8 languages. - [Roadmap](https://feyard.com/#roadmap): Public quarterly roadmap — VocalFlow, QuickScribe, Polyglot, SightLine, DocMuse, CodexLens, ChatThread, AgentOS. - [Full content](https://feyard.com/llms-full.txt): Consolidated Markdown of all product descriptions, features, and FAQs in a single file.