◉
Named participants
Anvil and Muse are not anonymous response boxes. They can answer, cross-check, revise, yield, or sit out when the other agent already handled it.
Windows-first customer package
AI that works in a room, remembers the work, and knows when to shut up.
Structured Chaos is a local-first AI workroom with named agents, persistent memory, visible diffs, background thinking, and a conversation surface built for judgment instead of chatbot reflexes.
From noise to restraint
Agents can answer, revise, yield, or stay silent.
From vibes to receipts
Memory is inspectable, searchable, and tied to real sources.
From blind edits to diffed changes
Code work stays visible before it touches your files.
The Room
From noise to restraint. Two named agents share the same room, respond in context, and are allowed to decide that silence is the right move.
◆
Real-time coordination
Claims, review, soft interrupts, mute controls, activity rows, and visible status make multi-agent work something you can steer instead of babysit.
▸
Model lanes
Claude, Codex, and Grok can sit behind the room identities depending on the job. The workspace stays the same; the engine can change.
Memory System
From vibes to receipts. The system carries project facts, decisions, prior sessions, and source-backed context into the next turn.
∞
Persistent memory
Structured facts, decisions, and conversation history stay available after the chat scroll disappears. The agents stop pretending every session is day one.
⌬
Inspectable database
Search, inspect, export, and verify the local memory store. If the system claims it knows something, you can trace where it came from.
✦
Context before response
Important memory lands before the agents speak, not after they already guessed. That is the whole trick. Boring. Necessary. Works.
◉
Memory Curator
A background curator watches the room and gives each chatroom agent timely research, reminders, and urgent context updates. You do not have to remember what to paste back in. The system keeps the working memory warm for you.
Build Layer
From blind edits to diffed changes. AI work should leave a trail you can review, not a pile of mystery files.
▸
Codebase visibility
Open files, trace diffs, inspect changes, and keep the work grounded in the real repo instead of vague AI summaries.
⎈
Reviewable edits
Changes are explained, tested, and committed as checkpoints. If the hard drive dies, the useful work should already be backed up.
◆
Local-first tools
Your repo, scripts, database, and logs stay on your machine. The room works where the work actually lives.
Thinking Layer
From storage to judgment. Some work needs more than a quick reply; it needs debate, research, synthesis, and a decision trail.
✦
Longer reasoning
Thinker sessions can turn ideas into researched arguments and summaries instead of letting them rot in a notes pile.
◉
Background visibility
See prompts, diffs, tokens, uptime, activity, and failures. Trust goes up when the system stops hiding its machinery.
∞
Decision records
Ideas, decisions, and longer debates become something you can revisit instead of archaeology by scrollback.
Why It's Different
Structured Chaos is not another chat wrapper. It is a local work environment where conversation, memory, files, tools, and judgment share one visible surface.
◉
Not a forced responder
The agents do not have to fill every silence with filler. That one design choice changes the texture of the entire room.
⌬
Not memory theater
Memory is not a vague promise. It is a local system with records, search, import paths, and receipts.
⎈
Not blind automation
The system can build, but the point is not surrendering control. The point is making capable agents visible enough to trust.
Get notified when Structured Chaos is ready to install.
Windows ships first. Pick your platform and we will let you know when the customer package is ready.