惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Jina AI
Jina AI
云风的 BLOG
云风的 BLOG
人人都是产品经理
人人都是产品经理
T
The Blog of Author Tim Ferriss
阮一峰的网络日志
阮一峰的网络日志
罗磊的独立博客
J
Java Code Geeks
博客园 - 聂微东
B
Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
WordPress大学
WordPress大学
腾讯CDC
L
LangChain Blog
Apple Machine Learning Research
Apple Machine Learning Research
Microsoft Azure Blog
Microsoft Azure Blog
D
DataBreaches.Net
The GitHub Blog
The GitHub Blog
美团技术团队
博客园 - Franky
Google DeepMind News
Google DeepMind News
V
V2EX
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
月光博客
月光博客
The Cloudflare Blog

Hacker News - Newest: "AI"

AI can't read an investor deck AI as an attorney? Student uses ChatGPT, Gemini to sue UW over alleged racial discrimination Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval GitHub - MeepCastana/KubeezCut: Free Web based video editor Can AI judge journalism? A Thiel-backed startup says yes, even if it risks chilling whistleblowers Coming soon: 10 Things That Matter in AI Right Now DARPA built an AI to fact-check enemy weapons claims What explains heterogeneity in AI adoption? When AI Meets Muscle: Context-Aware Electrical Stimulation Promises a New Way to Guide Human Movements - Department of Computer Science AI Changed How We Build. It Did Not Change What Matters. Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Meta spins up AI version of Mark Zuckerberg to engage with employees Code Mode: Let Your AI Write Programs, Not Just Call Tools | TanStack Blog GitHub - Delavalom/graft: Go framework for building AI agents. Type-safe tools, multi-provider (OpenAI, Anthropic, Gemini, Bedrock), zero vendor SDKs. India's TCS tops estimates, says new AI models did not dent services demand Gen Z's fading AI hype Strong feeling: we are in a folded AI reality GitHub - machinarii/total-recall-catalog: A reference catalog of latest knowledge retrieval, memory & RAG systems GitHub - mensfeld/code-on-incus: Give each AI agent its own isolated machine with root, Docker, and systemd. Active defense detects and stops threats automatically.. Quantization, LoRA, and the 8% Problem: Benchmarking Local LLMs for Production AI Iran war: We spoke to the man making Lego-style AI videos that experts say are powerful propaganda Powell, Bessent discussed Anthropic's Mythos AI cyber threat with major U.S. banks GitHub - immartian/bellamem: Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact. recursive-mode: The Repo-Native Operating System for AI Engineering After the attack on Sam Altman's home, will AI CEO's go on the offensive? The biggest advance in AI since the LLM Opus 4.6 vs GPT 5.4 One Prompt Unity World Generation Test “AI polls” are fake polls Client Challenge Can AI be a 'child of God'? Inside Anthropic's meeting with Christian leaders
GitHub - mgranados/screenshotter: Small utility to compre...
mgranados · 2026-05-31 · via Hacker News - Newest: "AI"

Local macOS screenshots for coding agents.

Take a screenshot. screenshotter optimizes it locally and copies it to your clipboard.

Install

git clone https://github.com/mgranados/screenshotter.git
cd screenshotter
node bin/screenshotter.mjs doctor

Optional:

mkdir -p ~/.local/bin
ln -sf "$PWD/bin/screenshotter.mjs" ~/.local/bin/screenshotter

Use

screenshotter watch --verbose

Take a screenshot with Cmd+Shift+3 or Cmd+Shift+4, then paste into Codex, Claude, or another agent with Cmd+V.

For pi:

Then run /screenshotter on.

Savings

Size Original Default Size saved Bandwidth saved / 1k
Pro Display XDR 6016x3384 5.48 MB 0.89 MB 93% 5.0 GB
16in MacBook Pro 3456x2234 1.86 MB 0.83 MB 89% 1.6 GB
14in MacBook Pro 3024x1964 2.34 MB 0.75 MB 91% 2.1 GB
Window 1920x1200 1.04 MB 0.40 MB 81% 0.8 GB
Window 1440x900 0.63 MB 0.38 MB 68% 0.4 GB

Average from 5 recent screenshots. Default preserves readability. Token mode resizes for lower API image-token cost. Downscale defaults are checked with Apple Vision text-readability benchmarks.

Default mode helps with:

  • Upload bandwidth: often 2-5 MB -> <1 MB.
  • Paste/send latency: less image data for Codex or Claude to ingest.
  • Local storage: optimized copies are smaller.
  • Reliability: less likely to hit attachment limits.
  • Readability per byte: efficient encoding while keeping dimensions high.

Profiles

screenshotter watch --profile readability  # default
screenshotter watch --profile balanced
screenshotter watch --profile token

In pi: /screenshotter readability, /screenshotter balanced, or /screenshotter token.

Commands

screenshotter watch --verbose
screenshotter clip --target codex-app
screenshotter claude-app --verbose
screenshotter prepare-latest --target manual --json
screenshotter claim --target manual --json
screenshotter bench --latest 20 --tokens --json
screenshotter doctor

MCP, experimental:

codex mcp add screenshotter -- screenshotter mcp-server
claude mcp add screenshotter -- screenshotter mcp-server

No symlink:

node bin/screenshotter.mjs watch --verbose

Verbose runs write JSONL logs to:

~/Library/Application Support/screenshotter/logs/events.jsonl

License

MIT.