T3 Code from @theo, @jullerino & co is getting so much more powerful!
Didn't realise you can now create webhooks that can be called anywhere to trigger new threads.
This unlocks so many cool use cases like linear issues, support tickets, fixing CI/CD runs etc
🔥🔥🔥
WERE BRINGING CODEX COMPUTER USE TO HERMES.
If you haven’t used codex computer use yet, it’s 2-4x faster and cheaper than regular browser control.
And we’re reverse engineering it for Hermes…
this is Alan’s way!
WEEKEND PROJECT 🥽
what if you could play Valve’s Half-Life: Alyx on the Apple Vision Pro?
without a pc, without streaming, running natively on the headset itself?
A new Cadu TestFlight update is out 🎉
This is a big one. Cadu's core is completely rebuilt: a new API layer, new business logic and a new UI foundation, with most components reworked along the way. With changes this deep, it might happen that you hit a few new bugs, sorry for that.
Thanks everyone for testing ❤️ and please keep the feedback coming!
Release Notes:
🖥️ Multiple servers
• Stay connected to all your Hermes servers at once
• Sessions, bots and rooms from every server in one place
• Notifications, sharing, compose and Siri work across servers
• Smarter reconnects and clear notices for outdated servers
💬 Chat
• Chats open instantly at the latest message, without jumps
• Long conversations stream faster and stay smooth
• Per-chat model and thinking level
• Drafts and queued messages are kept
• Message timestamps
• Answer sudo and secret prompts right in the chat
🔔 Notifications
• Push for every question your agent is waiting on
• One Live Activity across all servers
🤖 Bots & rooms
• Bots show as working during Kanban and background turns
• Long rooms open at the newest message and load reliably
• Bot room limitations have been removed
🔑 Sign-in
• Every sign-in method available everywhere
• Paste any SSH key, with passphrases and RSA SHA-2
🛠️ Tools & plugins
• New Git changes view with diffs
• Mail works with Himalaya 2 and installs from the inbox
• More reliable plugin installs, now on Windows too
• Voice messages resume after a lock or a call
• Better iPad navigation
🎨 Look & feel
• Five orb app icons and orb avatars
• Halloween characters 🎃
• Ready for translation
⚡ Under the hood
• Rebuilt on a generated, typed Hermes API client
• Chat UI and design system split into their own packages
• Big performance upgrades across the app
GLM-5.3-Flash EXL3 for DGX Sparks is now Abliterated 👀
- A simple flag to enable/disable it.
- Uses different weights.
- Gated - you must agree to the terms.
- Only slight loss in performance, around 6.5%
Also in this release:
- Next turn in chat is ~1s, down from ~16s
- Parallel agents no longer overwrite memory.
- Other improvements and fixes
Get it here:
https://t.co/fQzcacQtu5