SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

Arcanum: Open-Source .NET 10 Local AI Inference Hub Launches as Single Native Binary

0
·2 views

A developer has publicly released the README for Arcanum, a local-first AI inference platform built on .NET 10 and distributed as a single self-contained native binary requiring no runtime installation. The tool supports multiple OpenAI-compatible API providers — including DeepSeek, Groq, and Ollama — as well as local GGUF models, with no vendor lock-in. Arcanum features SQLCipher-encrypted local storage, per-sandbox resource limits enforced via cgroups v2, versioned prompt workflows, and goal-driven autonomous agents with cross-agent delegation. It is fully compatible with existing OpenAI clients via standard endpoints and includes a testing layer to validate prompts and agent plans before deployment. The project is aimed at senior engineers and targets use cases where infrastructure control, data privacy, and auditability are priorities.

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)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

How a single log marker replaced fragile inference in a multi-site maintenance UI

A development team building a multi-site maintenance screen needed to highlight the currently processing site in real time using streaming backend logs. Their first approach inferred site status by watching for name changes in the logs, but non-monotonic log ordering caused sites to be falsely marked as completed before they even started. A second fix attempted to filter backward index jumps in the planned execution order, but init-phase log lines still triggered false state transitions. The team ultimately resolved the issue by anchoring state changes to a single explicit marker line, such as '[Site A] Starting maintenance', ignoring all other log output for inference purposes. A final bug involving over-broad state resets was also fixed by scoping completion clears only to sites in the current run, leaving unrelated recently-maintained sites unaffected.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Mac App That Adds Lo-Fi Effects to Any Music You Already Play

A developer created Lo-fi Anything, a Mac application that applies lo-fi audio effects — such as tape saturation, pitch drag, reverb, and vinyl crackle — to any app's audio output in real time. The tool was built out of frustration with having to switch between personal playlists and dedicated lo-fi study streams just to achieve a warm, mellow sound. Unlike existing solutions, it does not require installing audio drivers or setting up virtual devices; users simply select an app and choose from six effect presets. The app is available for a one-time payment of five dollars, includes a three-day free trial, and requires no account registration.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Web Developer Ditches Cloud Converters for Local Browser Tools After NDA Concerns

A web developer who regularly converts images for client projects switched away from cloud-based tools like CloudConvert and Convertio after a client contract introduced an NDA clause restricting third-party data processing. Uploading files to external servers posed a legal risk, prompting the shift to browser-based converters that handle all processing locally on the user's device. Tools like webp2png.io use the browser's Canvas API to convert images without the file ever leaving the machine, supporting batch conversions of up to 50 files while preserving transparency. Beyond privacy compliance, the local approach also proved faster — a 5MB WebP file converts to PNG in under two seconds with no upload queue or server wait times. The developer advises freelancers and agencies handling client assets to review converter privacy policies carefully, warning against any tool that claims uploaded content may be used to improve its services.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

CLAP Audio Encoder Tested on ESC-50 Environmental Sounds via ONNX Runtime

A developer exported the CLAP audio encoder from the laion/clap-htsat-unfused model to ONNX format and evaluated it on the ESC-50 dataset, which contains 2,000 five-second environmental sound recordings across 50 classes. The experiment used a fixed one-nearest-neighbor approach, selecting two clips per class for a total of 100 clips, with no classifier training involved. Embeddings were compared using cosine similarity to assess whether recordings from the same class clustered together and whether sounds aligned with broader category groupings. The pipeline ran on an Apple Silicon CPU using PyTorch 2.8.0 and ONNX Runtime, producing normalized 512-dimensional embeddings per audio clip. The full code and reproducible environment are publicly available in the kiarina/labs repository on GitHub.

0 comments Read more at DEV Community