8 min read電子帳簿保存法インボイス制度OCR日本
領収書・請求書をブラウザだけで整理する：電子帳簿保存法とインボイス制度の実務
電子取引データの電子保存義務、検索要件、インボイス制度の保存事項を前提に、領収書や請求書をクラウドに上げずに整理する実務手順。ローカルOCRで何ができて、何ができないか（タイムスタンプ等）も正直に整理します。
Read article →Guides & engineering notes
The Feyard Blog
How browser-based AI actually works, and how to get real work done with it — privately, for free, without uploading a single file. Looking for a finished job instead of an explanation? Start with the end-to-end workflows.
9 min readEnglish originalPDFSummarizationTutorialLocal LLM
Summarize a PDF Without Uploading It: How Local Document AI Works
A 60-page paper or a 200-page contract, summarized on your own machine. How text extraction, 8,000-character chunking, and map-reduce summarization fit together — plus why scanned PDFs come back empty and what to do about it.
Read article →9 min readEnglish originalBackground RemovalMattingTutorialWebGPU
Remove Image Backgrounds in Your Browser: How the Models Actually Work
Two segmentation models, a fixed 512×512 graph, and a licensing decision most tools never mention. How in-browser background removal works, which quality tier to pick, and how to get clean edges on hair, fur and glass.
Read article →9 min readEnglish originalOCRTutorialTesseractPrivacy
Extract Text from Images in Your Browser: The Complete OCR Guide
Turn screenshots, scans and photos into editable text without uploading them anywhere. How Tesseract.js works in WebAssembly, which page-segmentation mode to pick, when preprocessing helps, and how to audit accuracy using confidence scores.
Read article →6 min readEnglish originalImage CompressionTutorialWeb Performance
How to Compress Images Without Uploading Them Anywhere
A practical guide to shrinking JPEG, PNG, WebP, and AVIF files in your browser with the same codecs cloud services use — plus settings that balance file size and quality, with zero uploads.
Read article →6 min readEnglish originalSpeech-to-TextTutorialWhisper
How to Transcribe Audio Locally: A Private Speech-to-Text Guide
A complete walkthrough of transcribing meetings, interviews, and voice memos in your browser with Whisper-based speech-to-text — no uploads, no per-minute fees, with tips for accuracy and export.
Read article →8 min readEnglish originalWebGPUWebAssemblyEngineering
WebGPU vs WASM: How On-Device AI Actually Works in 2026
A plain-English tour of the two browser technologies that make local AI possible: WebAssembly for portable compute, WebGPU for GPU acceleration — what each does, when they fall back, and what to expect on your device.
Read article →7 min readEnglish originalPrivacyLocal AIWebGPU
Why Your AI Tools Should Run in Your Browser, Not the Cloud
Cloud AI means uploading your files to someone else's server. Browser-based AI flips that model: the model downloads once, then everything runs on your device. Here is why that matters for privacy, cost, and speed.
Read article →