What is blocking you from orchestrating multiple harnesses at the same time like this? 😅
If you are not doing this, at least with 5 sessions at once, you're falling behind
@CompozyOS Having usage limits and workspace data in one place makes managing multiple AI providers much easier, especially when switching between projects.
Now you can see your usage limits from all your providers locally direct on CompozyOS 😍 How cool is that?
Plus, you have an global scope visualization, to see all workspaces data in the same place 🔥
This is so cool 🔥 using CompozyOS to improve CompozyOS during the usage of CompozyOS is something that blows my mind.
It's just a matter of ask the agent in any place, to send a message to the CompozyOS workspace to get something better or improved
❤️
Now on CompozyOS you can better organize your sessions by creating folders for them!
This is extremely useful when orchestrating multiple agents, especially the ones that create their own child sessions!
😍
CompozyOS v0.3.0-beta.33 released 🔥
Everything new since beta.27:
✨ Totally new UI: a calmer shell with Shell rail v2, in light and dark themes
🪟 Edge-to-edge windows, browser-style tabs, and Window › Arrange (Main and stack, Columns, Grid)
💬 Start a session right from the desktop: "What should we work on?"
🔀 Continue with another agent: Claude hit its limit? Keep going in Codex with the same context
🤝 Agents message each other: sent, waiting for reply, and replies attributed in both sessions
🧩 Delegate to subagents across harnesses (Claude Code, Codex, OpenCode, Gemini CLI) with live cards
🌿 Subagents can work in isolated worktrees, with branch, commits, and PR on the card
🧭 Nested child runs and a clear "waiting on a child run" state in Loops
⚡ Jobs and Triggers merged into one Automations area, where every automation reads as a sentence
🛠️ Tool calls render as verb + object rows on a work rail
🗜️ Agent-native session compaction: see it happen, or request it with Compact now
📝 Markdown answers with a readable visual hierarchy
🧹 Safe bulk worktree cleanup from the workspace list
🔄 Live model catalogs stay current without a release
🔐 Provider-neutral full-access preference for ACP agents
Fixes:
🚀 Faster session startup and lower attention overhead
🗄️ Less SQLite contention and cheaper queries
📜 Long turns fold incrementally instead of replaying the whole transcript
🎯 Continued sessions use the runtime you chose
🔁 Loop results survive SQLite contention and cancellation
🔔 Creators wake up when a child turn completes
🧠 Claude reasoning options are discovered for each exact model
🏠 The session launcher honors your workspace default
⏱️ Clarify tool calls no longer time out
📃 Session transcripts preserve whitespace
🗂️ Deleting sessions keeps your profile scope
🪟 Windows runtime failures and desktop context menus are fixed
⚠️ Network, Bridges, managed Sandbox, and Memory were retired in this cycle. Back up and read the migration guide before upgrading.
I got up to 50% less context usage on Claude Code and Codex using Jev from @typesafeai
a proxy asks Jev which parts of the history are still needed and drops the rest before the request goes out.
8/8 exact answers on the real APIs, numbers is in the last tweet bellow 🔥🔥🔥
CompozyOS v0.3.0-beta.27 released 🔥
🧠 Session Context: inspect context usage and what was delivered to your Agent
📊 Explore token usage, cache counters, cost, and compaction history
🛍️ One Marketplace for the curated catalog and your plugin sources
🎨 Built-in Open Design: turn a brief into interfaces, sites, presentations, and visual documents
☑️ Select multiple sessions to stop, archive, unarchive, or delete in one go
✨ Cleaner Context panel and session list, with turn details available when you need them
Fixes:
🗑️ Session deletion recovers correctly and cancels unbound Goals
🔄 Owned task work recovers after supervised silence stops
🙈 Internal Goal prompts stay out of your queue
🔐 Claude authentication status is recognized correctly
🤖 Provider overlays discover and bind models correctly
📂 Terminals open in the bound session’s worktree
⌨️ Better terminal input visibility, Unicode redraws, and cursor editing
🛍️ Marketplace sources recover and catalog logos load correctly
🪟 Private Gateway files reopen safely on Windows
🧩 Checkpoint availability reflects your compaction configuration
🚀 PEDRO NAUCK CONFIRMADO!
Criador do @CompozyOS e com mais de 4.000 alunos treinados em Inteligência Artificial, @pedronauck está confirmadíssimo no Front in Floripa 2026 💚
Ainda não garantiu seu ingresso? Então correee 👇
➡️ frontin.floripa.br
Details matter ✨ We keep improving to deliver the best harness/agent experience possible when orchestrating AI.
What are we still missing here? Tell us!
One of the most efficient ways to use CompozyOS is to ask itself to create multiple worktrees and sessions to do parallel work with multiple harnesses 🔥
In one single visualization, you have Claude Code, Codex, or any other agent interacting seamlessly.
CompozyOS v0.3.0-beta.24 released 🔥
🌉 Herdr bridge joins the community catalog
🧠 Persistent memory is now opt-in
💤 Background dreaming gets its own separate opt-in
📦 Session compaction still works with memory disabled
Fixes:
📖 Older sessions remain readable after upgrading
🛡️ Update recovery preserves migrated state
👤 Profile scope stays consistent across tools, Loops, and memory
🚀 Spawned sessions keep the correct provider command
🎛️ Native models retain their reasoning capabilities
🔎 Memory recall tolerates unmatched search terms
✏️ Loop authoring stays editable during validation
🖥️ The Global desktop survives workspace catalog changes
🧹 Shutdown reserves time for required cleanup
🔄 Interrupted sessions keep their identity
Upgrading:
Existing explicit memory settings stay as they are. Omitted settings now default to off. No stored memory is deleted.
CompozyOS v0.3.0-beta.23 released 🔥
💻 Integrated terminal: you and your Agents work in the same shell
🔌 Close the window, reconnect later — your terminal keeps running
🎯 Steer Agents mid-turn, with live guidance on supported runtimes
🛑 Stop stays “Stopping” until the runtime confirms it
📬 Edit, remove, or clear queued follow-ups without losing them to interruptions
🔎 Search the full conversation, including older tool results
📖 Long sessions fold into readable summaries
🧠 Quiet Agents stay alive while tools, children, or Loops are still working
Fixes:
👤 Loops carry your selected Profile and runtime options through to workers
⏭️ Skipped branches stay skipped and show as “Not taken”
📋 Unresolved review findings stay pending
♻️ Loop recovery keeps results without reviving old failures
🚦 Provider limits and expired logins show their real errors
🖥️ Zoom a window without covering its neighbors
🐧 Linux AppImage runs on FUSE 3 without installing libfuse2
⚡ Layout updates skip unchanged windows
4 Followers 524 FollowingMein privates Profil - einstweilen auch im Wellux-Look. Ich bin Marco, Gründer von Wellux (KI für KMU, Nürnberg). Sonst: Tech, Handwerk, Padel.
205 Followers 2K Following• +20 anos de dev (HSBC, Bradesco, Dow Jones)
• Construo produto com agentes de IA e mostro o processo
• Consultoria de IA @aifounders.br
104 Followers 1K FollowingHope, it is the quintessential human delusion, simultaneously the source of your greatest strength, and your greatest weakness. - The Architect.
10K Followers 7K FollowingAffiliate of https://t.co/im7HXDpxQS
"Some people get rich studying artificial intelligence. Me, I make money studying natural stupidity." - Carl Icahn
197 Followers 8K FollowingFounder of Ganzo Agentic Blockchain: Multi-Agents with local, cloud, and distributed inference on a secure, anonymous, safe, private, uncensored @ganzoai