Developer tools, frameworks, APIs, and tutorials — 1726 articles
Hugging Face · May 9
OncoAgent, an open-source clinical decision support system for oncology, routes patient queries to fine-tuned 9B or 27B LLMs based on complexity, enforces strict privacy controls, and fine-tunes on 26
OpenCode · May 9
**Scout agent** enables automated repo research, docs lookup, and dependency inspection within OpenCode workflows.
AI Jason · May 9
Developer documents common `/goals` command misuses in Claude Code and shares working patterns learned from trial and error—concrete gotchas like over-scoping goals, bad formatting, and when to break
Towards Data Science · May 9
Engineer discovered RAG systems can't distinguish current from outdated information and built a temporal filtering layer to fix it—turning a critical bug in an AI tutor into a reusable pattern for kno
Dev.to · May 9
Developer built a local AI coding agent on M5 Max (128GB) that failed 164 times before passing 35 tests, enforcing test-driven development throughout. Post documents the iteration cycle and constraint
Dev.to · May 9
Developer benchmarked 5 LLMs across 10 real-world agent coding scenarios and found a free model outperformed paid competitors. Concrete task results and cost analysis included.
Brave Discovery · May 9
Amazon backtracked on forcing employees to use its in-house Kiro coding tool, now allowing access to OpenAI's Codex and Claude via Bedrock after internal pushback. The reversal exposes how far Amazon'
MarkTechPost · May 9
GitHub open-sourced Spec-Kit, a toolkit that flips how developers use AI coding agents by requiring formal specifications before generation rather than vague prompts—addressing the 'vibe-coding' probl
Dev.to · May 9
Engineer with GCP credits set up a local gateway to run Claude Code via Vertex AI's Claude models instead of paying for direct Anthropic API access. Shows how to reuse existing cloud budgets with AI c
MarkTechPost · May 8
OpenAI released a Chrome extension for Codex that lets its AI agent interact with signed-in websites like Gmail, Salesforce, and LinkedIn directly in the browser, bridging the gap between sandboxed in
Simon Willison · May 8
Anthropic's Thariq Shihipar makes the case for requesting HTML outputs from Claude instead of Markdown, showing how Claude can render SVG diagrams, interactive widgets, and styled visualizations that
DZone AI · May 8
The gap between flashy LinkedIn AI agent demos and production systems is growing in 2026. While any developer can build a working prototype with Claude or GPT-4 APIs, shipping agents that handle real-
GitHub Copilot CLI · May 8
Slash commands can now appear mid-input, and multiple skills can be invoked in a single message.