🚨 BREAKING: China just dropped an AI employee that works 24/7 — and it's 100% open source.
Meet DeerFlow.
It researches, writes code, builds websites, creates slide decks, generates videos, and executes tasks entirely on your computer.
Not a chatbot.
Not an AI copilot.
An AI worker with its own computer.
Here's what makes it wild:
• Creates a plan and spins up multiple sub-agents in parallel
• Writes code, tests it, fixes errors, and keeps going autonomously
• Builds websites, reports, dashboards, and presentations from scratch
• Learns your preferences and improves over time
• Reads your files and delivers finished outputs, not drafts
• Searches the web, runs commands, and uses external tools
One prompt → Lead agent plans → Specialized agents fan out → Results merge → Finished deliverable.
A single task can launch dozens of agents working simultaneously while you sleep.
The craziest part?
DeerFlow 2.0 was completely rebuilt from scratch after users pushed it far beyond its original purpose.
Now it has:
⭐ 22.7K GitHub stars
🔀 2.7K forks
🏆 #1 on GitHub Trending
🏢 Built by ByteDance
🆓 100% Open Source (MIT)
We're entering the era where AI doesn't just answer questions.
It does the work.
Kasih AI Agent kamu "mata" ke seluruh internet!
Agent-Reach baru viral di komunitas:
Satu CLI untuk baca & search:
• Twitter/X
• Reddit
• YouTube (subtitle)
• GitHub
• Bilibili
• Xiaohongshu
• + web biasa & semantic search
Fitur utama:
✅ Gratis total (zero API fee)
✅ Auto routing + failover (kalau satu backend kena blokir, otomatis ganti)
✅ Privacy-first (cookie disimpan lokal)
✅ Bisa install langsung lewat agent kamu
✅ Ada `agent-reach doctor` buat cek kesehatan semua platform
Cocok banget buat:
- Research agent
- Competitor analysis
- Content curation agent
- On-chain / crypto research
Repo: https://t.co/WVGX1ZgvLE
Mau coba? Tinggal bilang ke agent kamu:
"帮我安装 Agent Reach" atau "Install Agent-Reach for me"
Kamu biasanya pakai tool apa biar agent bisa akses web?
Ini tutorial bikin video YouTube pake AI yang menurut gue paling gampang buat diikutin
Tapi gue yakin kurang dari 0,00001% orang yang nonton beneran coba dan praktekin ini
We now support rich formatting for all chatbots.
Tables, nested lists, inline media, formulas, headers and more — right in Telegram messages.
🔨 Start building! Docs: https://t.co/zgzPOOUJF5
i open sourced a Hermes skill for grading SOUL.md files.
100pt rubric for the identity layer:
> mission
> role boundaries
> hard constraints
> authority + escalation
> truthfulness
> success artifacts
> runtime hygiene
repo in reply:
Stop “vibe coding” pake AI.
Mulai sekarang kamu bisa bikin AI agent kamu berpikir seperti Product Manager sungguhan.
Dengan PM Skills Marketplace, kamu tinggal pakai slash command:
`/discover` → full discovery process
`/write-prd` → langsung generate PRD lengkap
`/strategy` → buat Product Strategy Canvas
`/ship-check` → audit AI-built code sebelum launch
`/red-team-prd` → stress test asumsi & risiko
68 skills + 42 chained workflows yang encode framework PM dari Teresa Torres, Marty Cagan, dan lainnya.
Hasilnya bukan cuma dokumen cepat, tapi **keputusan produk yang lebih baik** + struktur yang bisa di-review.
Support Claude Code, Codex, Cowork, dan tools lain.
Repo: https://t.co/2kYO0GHiIu
Buat yang lagi shipping product pake AI (atau bantu orang lain shipping), ini salah satu toolkit paling powerful yang pernah ada.
Kamu lebih sering “vibe code” atau sudah mulai pakai struktur PM di workflow AI kamu? 👇
Google Antigravity es una locura.
Acaban de grabar un tutorial de 19 minutos sobre cómo construir este sitio web animado y galardonado con Antigravity + GPT Image 5.5
AI agent kamu nulis 80 baris code buat yang sebenarnya cuma butuh 1 baris?
Kenalan sama Ponytail plugin yang bikin AI coding agent berpikir seperti laziest senior dev di dunia.
"The best code is the code you never wrote."
Cara kerjanya:
1. Cek dulu: Apakah ini perlu? (YAGNI)
2. Sudah ada di stdlib?
3. Ada native feature di platform?
4. Sudah ada di dependency?
5. Bisa 1 baris?
6. Baru nulis code minimal
Hasil benchmark:
• 80–94% less code
• 47–77% cheaper
• 3–6× faster
Contoh:
❌ Agent nulis full date picker + library + wrapper + logic
✅ Ponytail: <input type="date"> <!-- ponytail: browser has one -->
Support: Hermes,Claude Code, Codex, Cursor, Antigravity, Pi Agent, OpenCode, dll.
Repo: https://t.co/EiEtlH22R9
Pasang sekarang & rasain bedanya. Agent kamu bakal jauh lebih "senior" (dan males nulis code berlebihan) 😂
Kamu sering kesal AI agent over-engineering?
Comment pengalamanmu 👇
HERMES AGENT WITHOUT TOOLS IS A CHATBOT.
WITH THEM IT BUILDS 3D TOWERS IN BLENDER,
CHECKS STOCK PRICES, AND DRIVES VS CODE.
@tonbistudio JUST DROPPED THE FULL GUIDE.
module 6 of his 10-part Hermes masterclass.
best breakdown of the tool layer anyone has published.
what you need to know:
TOOLS vs SKILLS vs MCP
skills = instructions (markdown, loaded into context)
tools = callable functions (Python, agent emits call, Hermes executes)
MCP = adapters to external systems (Blender, Stripe, Linear, Notion)
every tool has three parts:
→ the function (does the real work)
→ the schema (what the model sees to decide when to call it)
→ the registry (makes the tool exist in the agent)
the model never runs the function directly.
it emits a structured request (tool name + JSON args).
Hermes executes and returns the result.
if the tool fails, the error goes back as JSON.
the agent recovers instead of crashing.
TOOL SETS CONTROL THE SURFACE
hermes chat --tool-sets web
# only web tools loaded. no files, no terminal.
hermes chat --tool-sets safe
# read-only: web search, vision, image gen.
# no file writes. no terminal. no code execution.
mid-session:
/tools enable video
/tools disable terminal
or toggle in the dashboard:
hermes dashboard → Skills → Tool Sets
MCP SERVERS
two transport types:
STDIO (local subprocess) or HTTP (remote endpoint).
add a server:
hermes mcp
# or ask: "add the MCP server for Blender"
filter tools with include/exclude per server.
keep only what you trust.
security:
→ OAuth 2.1 PKCE (no long-lived tokens in config)
→ package scanning via api.osv. dev before launch
→ all MCP calls go through approval gates
HERMES AS MCP SERVER
hermes mcp serve
exposes 10 tools via FastMCP.
connect VS Code Copilot or Cursor
to your running Hermes instance.
BUILD YOUR OWN TOOL
He built a stock price tool live:
→ Python function calling Finnhub API
→ schema with name, description, parameters
→ registered in tool_sets.py
→ agent calls it automatically when relevant
any repeating API call in your workflow
can become a native tool.
full Hermes architecture deep-dive in the article 👇
Hermes Agent now has a production-grade WhatsApp Business Cloud integration: use it as a private WhatsApp bot for yourself or your team, or configure it for customer-facing support.
Connect an existing WhatsApp Business Cloud number or create one through Meta Business Manager, then run 'hermes whatsapp-cloud' to wire it into Hermes with guided setup, secure webhooks, media/voice support, read receipts, typing indicators, and interactive approval buttons.
Gilaa! Lagi explore AI tiba-tiba banget nemu website sedetail ini.
=>https://t.co/Vcx5aqxQwP
Kirain cuma list template doang. Ternyata ini marketplace lengkap buat komponen Claude Code.
Bukan cuma dokumentasi, tapi katalog yang beneran bisa dipakai. Seperti
• agent untuk security audit
• command generate test
• MCP untuk GitHub atau database
• skills
• hooks
• settings
• Bahkan loker untuk developer juga ada!
Jadi nggak perlu setup dari nol tiap minggu cuma buat eksperimen workflow baru.
Share ini buat yang lagi explore Claude Code biar nggak stuck di setup awal.
Note:
-This post is not an endorsement or sponsorship.
-Iam Just sharing in case it's useful to someone else.
ÇİNLİ BİR ADAM RESMEN PARA BASMA MAKİNESİ OLUŞTURDU.
Github'da 13.000 yıldız almış bir araç var.
Adı moneyprinterturbo.
Bir çinli geliştirici yaptı.
Ücretsiz ve tamamen açık kaynak.
Tiktok, reels, youtube shorts için tam videoları otomatik üretiyor.
Nasıl çalışıyor.
Tek bir iş akışında her şeyi hallediyor.
Senaryo üretimi, seslendirme, altyazı, görsel kaynaklar, düzenleme, hepsi dakikalar içinde.
Yayına hazır video çıkıyor.
Sen hiçbir şeye dokunmuyorsun.
Şimdi neden bu kadar popüler oldu.
Çünkü normalde bu süreç şöyle işliyor.
Senaryo için ayrı araç, seslendirme için ayrı araç, altyazı için ayrı araç, görsel için ayrı araç, düzenleme için ayrı araç.
Her biri ayrı para istiyor, ayrı zaman istiyor, ayrı öğrenme istiyor.
Moneyprinterturbo hepsini tek çatıda birleştirdi.
Ücretsiz, sınırsız ve kategorisindeki en popüler açık kaynak proje haline geldi.
Tiktok shop ve youtube shorts kanalları aylık 6 ila 10 bin dolar kazanıyor.
Bunlar bu süreci kullanıyor.
Fark şu:
Onlar araçlar için para ödüyor.
Sen ödemiyorsun.
Kurulumu 5 dakika.
Github'da ara.
Kur, çalıştır, içerik üret hepsi tamamen senin elinde.