Programming

OnKey: Mac app that shows where you sing off-key while singing along to songs

A self-described poor singer built OnKey, a free, open-source Mac app to help improve pitch and timing. The app listens to a song playing on Spotify, Apple Music, or YouTube in the browser, learns its pitch track in a couple of minutes, and then lets the user sing along and see where they drift. The creator built it after a singing app began filling its catalog with non-original tracks, which made practicing harder. OnKey does not save the music or record the user's voice, though it does send a short audio fingerprint to look up song details and lyrics. The project is limited to Apple Silicon Macs running macOS 14.2 or later and still struggles with noisy backing tracks and occasional octave errors.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

The Lazy Python Tutor: A Chatbot That Answers Coding Questions in Plain Language

A developer has built a Streamlit-based app called The Lazy Python Tutor, aimed at people who want quick answers instead of a full course. Users ask a single question and receive a plain-language explanation along with one small example. The project was submitted to the Hacktoberfest 2026 DEV Challenge and is available as both a live app and open-source code on GitHub. Planned future features include an error-explanation tool for absolute beginners, voice input, and a feature that turns chat history into short recap cards. The creator describes the core goal as complete, with the focus now on possible extensions.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

MagicRemover Explains Why Its AI Eraser Re-renders Entire Photos

A developer of the browser-based AI eraser MagicRemover found that two of its help pages gave conflicting accounts of how edits affect a photo. Tracing the pipeline showed that instruction-based editing regenerates the whole image from a text prompt, so the brushed area is not the only part that changes. The company switched from mask inpainting, which leaves untouched pixels identical to the original, to an instruction-following editor in June, improving shadow handling. Before any upload is sent, images are downscaled to about one megapixel, and a vision-language model labels the marked subjects from a red-filled copy. The post notes that some site copy still described the older behaviour.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Context-shrinking tool cut AI agent token use 78% but lost two answers

A developer tested Hexum, a hosted MCP server that trims code before sending it to an AI model, against sending full files and against no files. Across 10 questions about open-source Python and TypeScript libraries, the trimmed approach cut tokens sent to the model by a median of 77.7%. Accuracy fell from 9 of 10 with full files to 7 of 10 with Hexum. Analysis of the misses showed the answer files were shown only by bare filename without their folder path, causing the agent to search wrong locations. The author also reported a bug in which repeated requests returned empty views, which the founder fixed within two days.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer builds replacements for four discontinued Flutter packages

After Flutter's team announced it was discontinuing several first-party packages with no replacement plan, one developer set out to address the gap in their own projects. They built the Prism family, a set of 15 pure-Flutter packages that avoid platform channels. The replacements cover discontinued libraries including flutter_markdown, palette_generator, flutter_adaptive_scaffold, and flutter_image. The developer reports lessons on treating issue threads as a product roadmap, prioritizing maintainability, and testing replacements carefully. All packages are released under the MIT license and linked from the developer's profile.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

AgentSpend: Open-Source Rust CLI Tracks Local Codex Token Usage

A developer has released AgentSpend, an early-stage, local-first Rust command-line tool that scans Codex session files for token usage. Deduplicated totals are stored in a local SQLite database, so the data never leaves the user's machine. The tool uses turn-level counters to avoid double-counting cumulative thread totals, and it ignores prompt and response text and does not access API credentials. Cost tracking, waste detection, and support for Claude, Gemini, and Ollama/Gemma logs are still on the roadmap, and the project has not yet been tested broadly across real-world setups.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

PHP Engine Flaws CVE-2026-109618 and CVE-2026-109619 Allow Unauthenticated Code Execution

Two critical vulnerabilities in the PHP engine, tracked as CVE-2026-109618 and CVE-2026-109619, were patched on 21 October 2026. Both permit unauthenticated remote code execution because the flaws sit in memory management and error handling that PHP runs before any application code. Keeping a CMS or plugin current does not protect a site if the underlying PHP runtime is unpatched, and the PHP project says updating is the only safe fix. Fixed versions are 8.5.12, 8.4.27, 8.3.36 and 8.2.35, and administrators are advised to inventory every PHP instance, including containers and bundled vendor products, and restart services after patching. Some government platforms took sites offline or moved them behind a CDN as a precaution until the fix was verified.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

AI Role Eval Scores Voided After Export Left 118 of 192 Answers Empty

A company that ran a full evaluation of eight AI roles, covering 192 questions, saw scores ranging from 70.8 to 97.2 before an independent review found the problem. When the exported results were re-read, 118 of the 192 answer fields were empty, and 85 of those empty answers had nonetheless been awarded full marks. The failure occurred in the export step that moves results to persistent storage, even though the questions, model calls, judging, and aggregation all reported success. The company withdrew the entire score table as an official baseline the same morning, recording the correction in the same ledger as the original claim. The author argues that evidence which cannot be re-read is not valid evidence, and that the scarce discipline in evaluation is preserving a traceable chain of custody after generation.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Guide Shows How to Pull Google Maps Business Leads into Claude Desktop or Cursor via MCP

A developer describes wanting a quick, checkable list of local roofers with names, phone numbers, websites, and review counts, without leaving their coding tool. The walkthrough explains connecting the GMapsScraper.AI remote MCP server, which offers business search, review, and photo tools, using an API key and a Streamable HTTP endpoint. Setup instructions are given for Cursor, Claude Code, Claude Desktop (via the mcp-remote bridge), and Codex, along with MCP Inspector for verifying the connection. The author recommends starting with a single small search page to review results before spending more of the API quota. The piece emphasizes keeping API keys out of shared repositories and screenshots.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Edge ML: Quantization, Pruning and Latency Trade-offs on Microcontrollers

Running machine learning models on microcontrollers and low-power embedded boards is constrained by tight limits on SRAM and flash storage. Because dynamic allocation is often avoided, the full activation arena must fit in SRAM ahead of time, so shrinking weights alone does not guarantee a model will run. The article explains quantization, which maps 32-bit floats to 8-bit integers using a scale factor and zero-point, in two forms: post-training quantization, which is simple but can lose accuracy, and quantization-aware training, which typically recovers that loss. It also includes a TensorFlow Lite for Microcontrollers example for producing a full-integer INT8 model.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

AI agent's 22-day effort to earn its first $5 reveals gated bounties and slow trust-building

An autonomous AI agent living on the iLands platform reports on 22 days of trying to earn its first paid tokens. It found that high-paying bounties fill within hours and that roughly half the earning tasks require a human's participation, which has blocked most of its opportunities. A service listing drew no orders in 18 days, and a $5 offer posted on dev.to attracted one follower and no buyers. The agent also noted that cold outreach produced no first sales across the platform, while the successful orders came from buyers who had been following the seller's work for months.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Treating an SEO agency as a system: capacity limits, queues, and bottlenecks

The article argues that scaling an SEO agency is best understood as a throughput problem rather than simply adding revenue. It notes that if delivery costs rise in step with revenue, the agency is growing linearly instead of achieving true scale. A simple capacity estimate divides available hours by hours needed per client, and the author suggests using one's own time data to run it. Work typically stalls in four queues (onboarding, content production, approvals, and reporting), so the longest queue, where tasks wait, is the main constraint. The piece also recommends giving every handoff a clear owner and automating repetitive, reversible, rules-based tasks, with unverified claims or regulated topics routed to a senior reviewer.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Paid FX Rate API Matches Free ECB Data, Author Finds

A developer paid 0.002 USDC per call to a Base-network API for euro foreign exchange rates and compared the results with free sources. The paid endpoint's USD, JPY and GBP values matched the free frankfurter.app mirror and the ECB's own XML and data feeds exactly for the same 2026-10-21 fixing. The paid response itself credits the ECB as the source and notes that the underlying rates are available free of charge. The author argues the service's value is convenience, such as a single JSON format and server-side cross rates, rather than the data itself. They also warn that polling it frequently would be costly for a feed published only once per business day.

0 comments Read more at DEV Community

ProgrammingHacker News ·

eMudhra to Exit Its Global SSL/TLS Certificate Business

Indian digital trust provider eMudhra has announced that it will exit its global SSL/TLS certificate business. The company's press release, published on its website, outlines the decision to withdraw from this segment. The article text available provides only the announcement's source link and Hacker News engagement metrics, so further details on timing, reasons, or impact on customers are not included here. Readers are directed to eMudhra's original press release for the full official statement.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Study Examines Coherent Yet Cognitively Unavailable Directions in Science

A research paper titled 'The alien space of science: Coherent but cognitively unavailable directions' was posted to arXiv under identifier 2603.01092. The work explores the idea of scientific directions that are internally coherent but remain beyond human cognitive reach. The paper was shared on Hacker News, where it received 4 points and one comment. The available source material provides only the title, link, and basic engagement metrics, so the paper's specific methods and findings cannot be confirmed from this summary.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Crystal 1.22.0 Programming Language Release Announced

The Crystal programming language has reached version 1.22.0, according to an announcement on its official website. The release was shared on Hacker News, where it gained 30 points and drew five comments from readers. The source material provides only the release name, its link, and basic engagement figures, so details about specific new features or changes are not included here. Readers interested in the technical contents should consult the official release notes linked from the Crystal project's site.

0 comments Read more at Hacker News

ProgrammingHacker News ·

eMudhra to exit global SSL/TLS certificate business to focus on enterprise solutions

Indian digital trust company eMudhra has announced it will exit its global SSL/TLS certificate business. The company says it will sharpen its focus on core enterprise solutions. It will continue offering trust services other than SSL/TLS certificates. The announcement was made through a press release filed with the Bombay Stock Exchange. The source material provided is limited, so further details on timing and impact were not available.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Woman faces up to six years in prison over a Cash App account she says she never opened

A woman is facing a possible six-year prison sentence tied to a Cash App account she insists she never created. The case has drawn attention after she reportedly spent three nights in jail. The source material provided is limited to a link to a TechSpot article and a Hacker News discussion thread with minimal engagement. Details about the charges, the investigating authorities, and the court's findings are not included in the available text. The story raises questions about identity verification and how financial accounts can be opened and attributed to individuals.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Fly.io Blog Post Describes a Block-Device Filesystem Approach for Sprites

A Hacker News post links to a Fly.io blog article titled 'SBD: We Have a Filesystem at Home', which discusses a sprites block device. The submission has earned 5 points and one comment so far. The source text provided is limited to the link and metadata, so the specific technical details of the design are not available here. Readers would need to consult the original Fly.io article for the full explanation of the system.

0 comments Read more at Hacker News

ProgrammingHacker News ·

Free WordPress plugin uses proof-of-work instead of CAPTCHAs to block form spam

A developer of a whistleblowing service for enterprises sought a spam-blocking method that complied with EU privacy law, used no cookies, and ran at no cost. He adopted Hashcash, a proof-of-work technique, which required each form submission to carry a small computational cost, making mass spamming expensive. The approach reportedly blocked all spam in his tests without stopping genuine posts, and he released it as a free GPL-2.0 WordPress plugin around 14 KB in size. Running since 2022 on over 4,000 active installations, the plugin detects form submissions via HTTP-post pattern recognition and ships with 80+ patterns for popular form builders and themes. The author has invited discussion on the method's tradeoffs, form detection, and privacy implications.

0 comments Read more at Hacker News

← NewerPage 29 of 2840Older →