Tenant-Isolated Uploads: Scanning Private-Bucket Media Before Publication

·3 views

The article outlines a design for B2B SaaS products handling large media uploads, where bytes go straight into a private storage bucket and a durable event triggers thumbnailing and virus scanning. Tenant isolation is treated as an enforced authorization rule rather than a naming convention, with object keys built server-side from the authenticated session. Uploads are quarantined, and nothing is published until validation and scanning produce a durable verdict. The author also stresses replayable, idempotent processing so that delayed or duplicated notifications do not create duplicate or cross-tenant outputs. A boundary table in the article lists the required behavior and likely failure modes at each system interface.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Kisumu shops begin accepting bitcoin via QR codes that settle in M-Pesa

Two small shops in Kisumu started accepting bitcoin on 8 November 2026, letting customers pay by scanning a QR code with any Lightning wallet. The payments are converted so that shillings reach the merchant's M-Pesa account, meaning shopkeepers need no wallet, exchange account, or bitcoin knowledge. A local advocate is personally visiting shops to explain the system and printing a QR code for each merchant, at a cost of about $1 per code that he says he covers himself. He aims to add roughly one new merchant a week and is seeking sponsors and collaborators, while also preparing a Btrust application. The effort is intended to shift local perceptions of bitcoin, which many residents see as a scam or gamble.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Proposed Three-Layer AI Framework Aims to Make Clinical Decision Support More Transparent

A new article proposes a clinical decision support architecture that combines rule-based protocols, Bayesian reasoning, and time-series prediction, with physicians kept as final decision-makers. It argues that current systems often act as opaque black boxes, which fuels alert fatigue and clinician distrust. The authors recommend judging clinical usefulness through outcome-action pairing and decision-curve metrics rather than accuracy alone. The piece also reviews regulatory requirements, including the EU AI Act and FDA guidance classifying such software by risk, and it addresses ethical principles and bias mitigation.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

PicReady: a .NET library for resizing, cropping and compressing web uploads

The author built PicReady, a .NET 8/9 library that wraps SkiaSharp to handle common web-app image tasks such as resizing, cropping, avatar creation, WebP conversion and EXIF stripping. The library addresses problems with raw phone photos, including sideways orientation tags, large file sizes and the risk of decompression bombs on public upload endpoints. The author compares it with alternatives such as Magick.NET, ImageSharp and Six Labors' commercial Boutique license, which costs $799 per year for up to 10 developers. A free MIT-licensed subset, PicReady.Lite, offers basic resizing without a license key. Both packages target .NET 8 and .NET 9 and avoid System.Drawing and ImageSharp dependencies.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Audio Formats and Quality Settings Shape Voice AI Performance

The article explains that the audio format and quality settings chosen for voice AI output affect file size, latency, cross-platform compatibility, and how natural a synthesized voice sounds. It compares WAV, MP3, AAC, OGG Vorbis, and Opus, noting that WAV offers the purest signal but produces much larger files, while MP3 and AAC balance size and quality for most web and mobile uses. Opus is highlighted as well suited to low-latency real-time voice applications. Key parameters such as sample rate, bit depth, and bitrate are outlined along with typical values for text-to-speech. A Python example using the ElevenLabs API demonstrates how to request a specific output format, sample rate, and bitrate.

0 comments Read more at DEV Community