#devtools — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #devtools, aggregated by home.social.
-
🚀 llama.cpp b10436 is out
mtmd, common: various fixes (#27071) · * apply fixes · * cont · * revert gguf fix · · Website: - · macOS/iOS: - macOS Apple Silicon (arm64) -…
⭐ 123,916 stars
All open-source AI releases →https://olud.ai/releases.html
-
🚀 llama.cpp b10436 is out
mtmd, common: various fixes (#27071) · * apply fixes · * cont · * revert gguf fix · · Website: - · macOS/iOS: - macOS Apple Silicon (arm64) -…
⭐ 123,916 stars
All open-source AI releases →https://olud.ai/releases.html
-
💡 Open-source pick of the day: Wan2.1
Wan2.1 is a family of open video generation models from Alibaba covering text-to-video and image-to-video, wit…
⚡ Olud Pulse: 51/100 -
💡 Open-source pick of the day: Wan2.1
Wan2.1 is a family of open video generation models from Alibaba covering text-to-video and image-to-video, wit…
⚡ Olud Pulse: 51/100 -
🔧 The backend is #Python-based on top of #pipecat, the UI is #TypeScript. The architecture is modular so components can be swapped, and #Docker-first for consistent deployments. Python and Node SDKs are published on PyPI and npm.
📊 Licensed under BSD 2-Clause, ~4.2k stars, 856 forks, 42 releases. #MCP native, #selfhosted deployment with HTTPS documented, plus a managed cloud version. #devtools
-
🔧 The backend is #Python-based on top of #pipecat, the UI is #TypeScript. The architecture is modular so components can be swapped, and #Docker-first for consistent deployments. Python and Node SDKs are published on PyPI and npm.
📊 Licensed under BSD 2-Clause, ~4.2k stars, 856 forks, 42 releases. #MCP native, #selfhosted deployment with HTTPS documented, plus a managed cloud version. #devtools
-
How I tear apart code I did not write: start with the first thing that runs and follow it. Call datetime.api, get a datetime, fine. Get a street address, problem. Generated code is locally clean and globally inconsistent. Readable is not coherent.
#AI #SoftwareEngineering #CodeReview #BuildInPublic #DevTools
-
How I tear apart code I did not write: start with the first thing that runs and follow it. Call datetime.api, get a datetime, fine. Get a street address, problem. Generated code is locally clean and globally inconsistent. Readable is not coherent.
#AI #SoftwareEngineering #CodeReview #BuildInPublic #DevTools
-
🚀 CrewAI 1.15.16 is out
What's Changed · Features - Introduce execution context management with UUID support - Record what kind of exception ended a flow - Record when a…
⭐ 57,060 stars
All open-source AI releases →https://olud.ai/releases.html
-
🚀 CrewAI 1.15.16 is out
What's Changed · Features - Introduce execution context management with UUID support - Record what kind of exception ended a flow - Record when a…
⭐ 57,060 stars
All open-source AI releases →https://olud.ai/releases.html
-
Watching your LLM bill grow in real time 💀
🪙 **token-visualizer** — Simulate LLM conversations and visualize token usage & cost
💯 Compare pricing models and watch per-turn costs grow
🦀 Written in Rust & built with @ratatui_rs
-
Watching your LLM bill grow in real time 💀
🪙 **token-visualizer** — Simulate LLM conversations and visualize token usage & cost
💯 Compare pricing models and watch per-turn costs grow
🦀 Written in Rust & built with @ratatui_rs
-
Notable fork: Void reworks VS Code with built-in AI — tab autocomplete, inline edits, agent mode, your keys, full control. Currently scoring 58/100 on Olud Pulse, it’s worth seeing how it stacks up.
-
Notable fork: Void reworks VS Code with built-in AI — tab autocomplete, inline edits, agent mode, your keys, full control. Currently scoring 58/100 on Olud Pulse, it’s worth seeing how it stacks up.
-
Regex101 but in your terminal!
👀 **regexplain** — A TUI for explaining and visualizing regular expressions
💯 See what each part of your regex does and how it matches your input
🦀 Written in Rust & built with @ratatui_rs
⭐ GitHub: https://github.com/kapilpokhrel/regexplain
#rustlang #ratatui #tui #regex #terminal #devtools #opensource
-
Regex101 but in your terminal!
👀 **regexplain** — A TUI for explaining and visualizing regular expressions
💯 See what each part of your regex does and how it matches your input
🦀 Written in Rust & built with @ratatui_rs
⭐ GitHub: https://github.com/kapilpokhrel/regexplain
#rustlang #ratatui #tui #regex #terminal #devtools #opensource
-
🚀 Pydantic AI v2.29.0 is out
What's Changed 🚀 Features * Support FastMCP 4 and MCP SDK v2 in MCPToolset alongside FastMCP 3 by @adtyavrdhn in…
⭐ 19,260 stars
All open-source AI releases →https://olud.ai/releases.html
-
🚀 Pydantic AI v2.29.0 is out
What's Changed 🚀 Features * Support FastMCP 4 and MCP SDK v2 in MCPToolset alongside FastMCP 3 by @adtyavrdhn in…
⭐ 19,260 stars
All open-source AI releases →https://olud.ai/releases.html
-
DeepSeek, Alibaba, and xAI all shipped frontier models within 24 hours with benchmark scores neck and neck. Also: Cognition raising at 40B and Twitch opting streamers into AI training by default.
-
DeepSeek, Alibaba, and xAI all shipped frontier models within 24 hours with benchmark scores neck and neck. Also: Cognition raising at 40B and Twitch opting streamers into AI training by default.
-
- Auth0
- Clerk
- Firebase Auth
- Supabase AuthWhich one would you pick for a new app?
-
Blocked “aria-hidden”: The Warning Is Right, and Every Fix You’ve Found Is Wrong, by @csstricks:
https://css-tricks.com/blocked-aria-hidden-fix/?ref=frontenddogma.com
-
Blocked “aria-hidden”: The Warning Is Right, and Every Fix You’ve Found Is Wrong, by @csstricks:
https://css-tricks.com/blocked-aria-hidden-fix/?ref=frontenddogma.com
-
AI coding agents are amazing, but their context collapse is real. Stop re-explaining your entire codebase every session. 🔁
I open-sourced Vibe Coding OS to give Claude & Cursor persistent memory using Obsidian and MCP. It structures your markdown vault as a project brain, letting your agents read your architecture and write their own dev logs so they never forget. 🧠💻
Grab the repo here:
🔗 https://github.com/iabdullahahmad/VibeCodingOS#VibeCoding #Obsidian #AIAgents #DevTools #SoftwareEngineering
-
AI coding agents are amazing, but their context collapse is real. Stop re-explaining your entire codebase every session. 🔁
I open-sourced Vibe Coding OS to give Claude & Cursor persistent memory using Obsidian and MCP. It structures your markdown vault as a project brain, letting your agents read your architecture and write their own dev logs so they never forget. 🧠💻
Grab the repo here:
🔗 https://github.com/iabdullahahmad/VibeCodingOS#VibeCoding #Obsidian #AIAgents #DevTools #SoftwareEngineering
-
Found a task runner that changes everything!
🌀 **upmd** — Turn Markdown code blocks into runnable tasks and dependency-aware workflows
💯 Keep instructions, commands, dependencies, and live terminal output together in the same document
🦀 Written in Rust & built with @ratatui_rs
⭐ GitHub: https://github.com/rezigned/upmd
#rustlang #ratatui #tui #markdown #devtools #cli #opensource
-
Found a task runner that changes everything!
🌀 **upmd** — Turn Markdown code blocks into runnable tasks and dependency-aware workflows
💯 Keep instructions, commands, dependencies, and live terminal output together in the same document
🦀 Written in Rust & built with @ratatui_rs
⭐ GitHub: https://github.com/rezigned/upmd
#rustlang #ratatui #tui #markdown #devtools #cli #opensource
-
Docs always drifting out of sync with your code? 🧩
New open-source agent skill — docs-update-expert: reconciles README, CHANGELOG, AGENTS.md, API references & inline comments with your repo's actual state. It builds a git-based drift map, edits surgically, and never fabricates facts.
https://www.skills.sh/ricardodantas/skills/docs-update-expert
-
Docs always drifting out of sync with your code? 🧩
New open-source agent skill — docs-update-expert: reconciles README, CHANGELOG, AGENTS.md, API references & inline comments with your repo's actual state. It builds a git-based drift map, edits surgically, and never fabricates facts.
https://www.skills.sh/ricardodantas/skills/docs-update-expert
-
🚀 CrewAI 1.15.15 is out
What's Changed · Features - Report flow outcome, duration, and human-in-the-loop signals. · Bug Fixes - Emit FlowStartedEvent when a boundary hook…
⭐ 56,970 stars
All open-source AI releases →https://olud.ai/releases.html
-
🚀 CrewAI 1.15.15 is out
What's Changed · Features - Report flow outcome, duration, and human-in-the-loop signals. · Bug Fixes - Emit FlowStartedEvent when a boundary hook…
⭐ 56,970 stars
All open-source AI releases →https://olud.ai/releases.html
-
🚀 JJ-FZF 0.42.0 is out!
⚔️ Alt-L: interactive conflict resolution with jj resolve
🔄 Ctrl-R: rewrite change ID (jj metaedit --update-change-id)
📜 Ctrl-O: keep showing *all* oplog changes (jj-0.40+)
🧠 Cleaner LLM commit message generation
📝 Supports jj describe --editor (jj-0.42)
⏱️ Error pause delays configurable via $JJFZF_ERROR_DELAY
🔧 Fixed bookmark startup position & quoting⚠️ Requires jj-0.42.0, fzf-0.67.0, GNU Make ≥ 4.0, bash ≥ 5.1
🔗 Release:
https://github.com/tim-janik/jj-fzf/releases/tag/v0.42.0 -
🚀 JJ-FZF 0.42.0 is out!
⚔️ Alt-L: interactive conflict resolution with jj resolve
🔄 Ctrl-R: rewrite change ID (jj metaedit --update-change-id)
📜 Ctrl-O: keep showing *all* oplog changes (jj-0.40+)
🧠 Cleaner LLM commit message generation
📝 Supports jj describe --editor (jj-0.42)
⏱️ Error pause delays configurable via $JJFZF_ERROR_DELAY
🔧 Fixed bookmark startup position & quoting⚠️ Requires jj-0.42.0, fzf-0.67.0, GNU Make ≥ 4.0, bash ≥ 5.1
🔗 Release:
https://github.com/tim-janik/jj-fzf/releases/tag/v0.42.0 -
Nimbalyst v0.71.0 now runs Claude Code and Codex agents on macOS, Windows, and Linux from one workspace. The MIT-licensed desktop app lets users compare edits side-by-side, but still requires separate provider access for each agent. #AI #OpenSource #DevTools https://www.implicator.ai/nimbalyst-claude-code-codex-mac-windows-linux/
-
🔦 Open-source tool of the day: vLLM
vLLM is a high-throughput inference and serving engine using PagedAttention to maximize GPU utilization, the d…
⚡ Olud Pulse: 75/100 -
LLMs make it incredibly easy to outsource thinking, not just work. I keep wondering what happens if convenience slowly erodes our own cognitive ability.
With CodeMeridian I’m experimenting with the opposite direction: human-cognitive-seed.
human seed → AI expansion → human evaluation → improved human model
AI handles the drudgery, but for reasoning it should challenge, add context, preserve the “why”, and leave us understanding more than before.
-
Intelligent Terminal 0.2 is here with local model support
#terminal #windows #cli #ai #commandline #windowsdev #devtools
-
PowerToys Preview v0.101.2211.0 Release | GitHub
https://github.com/microsoft/PowerToys/releases/tag/v0.101.2211.0
#devtools #powertoys #windows #ai #phisilica #advancedpaste #cmdpal