You drop a thought
Two modes: deliberate (you add it in the Brain) or ambient (you drop it in chat and the intelligence picks it up). It captures more than ideas — it maps decisions and the knowledge that holds the business together.
An intelligence layer for your organization
DacoWork follows what happens, understands your intent, and proactively turns conversations, ideas, and decisions into concrete action — closing the open loops that today die in a meeting note, in a terminal about to shut down, or in your head. You approve.
An open loop is the concrete next step that would unblock things — sometimes depending on other people, sometimes only on you — and that never gets taken. It sits dying in a meeting note, a chat, your head, while the clock runs and the context that made that step easy evaporates with it.
The disease is the open loop at scale: meetings where much is discussed, action items stalled on whoever doesn't follow through, client fires, rework. DacoWork is an intelligence layer that was there in the session and keeps the context — it sees the loose end and, at the right moment, hands you the next step ready to fire. You just approve.
“The intelligence was there in the session and keeps the context. It doesn't die when you close your laptop.”
It's not a list of tools. It's an intelligence layer — a brain and a voice — that knows your business and acts at the edge of the work, at the level of proactivity you authorize.
Two modes: deliberate (you add it in the Brain) or ambient (you drop it in chat and the intelligence picks it up). It captures more than ideas — it maps decisions and the knowledge that holds the business together.
A repository of organizational intelligence — semantic, temporal, multi-tenant. A passive substrate: it keeps the context and exposes it to whoever has permission. The business's memory lives here, not in a vendor's product.
It classifies each item against the business's intent and calibrates how proactive to be — from merely suggesting, to preparing, to acting with your approval. This is where pre-authorized decisions and the human-in-the-loop live.
The workspace's voice closes the loop: the email that unblocks, the action items from the notes, the script for Monday's meeting — ready to fire. It calls in a specialist when expertise is needed. The initiative is real; the approval is yours.
Monitoring surfaces show what the intelligence saw, decided and did — per project and across the whole workspace, scoped by your permission.
Channels feed it. The agents — yours, ours, your clients' — operate on it through a single MCP gateway. No integration tax, because the business's memory is the common ground.
Open Brain is a Postgres store, a set of credentials, a hybrid search. Every agent is a tenant.
A captured signal becomes a next step ready for you to approve — no status meeting afterward. Today, with you in the loop. The fully autonomous edge is what we're building.
Track A — what ships today, human in the loop. Autonomous action at the edge is Track B, under construction.
From the small organization to the enterprise. The focus is transforming the legacy organization: moving from meetings to re-share context → to an intelligence layer that filters everything, maps the action items, and ensures execution.
Writes the principles the brain applies. Their intent is the filter the intelligence triages everything through.
Reads what the intelligence prepared, adjusts, approves and delivers. Looks after the business with the intelligence layer, not despite it.
The organization operates through its clients' sub-workspaces, scoped by permission. White-label, made concrete.
You see what was captured, decided and done — per project and across the whole workspace, scoped by your permission.
A project's dashboard: activity feed, heartbeats, severity routing. Shipped.
Everything happening in the workspace, scoped by permission, across clients' sub-workspaces. Under construction.
The brain exposed as MCP: each person's agent plugs in and queries the business's memory.
Project Channel ships today. The cross-sub-workspace Workspace Feed is the target — the permission-scoped assembly is under construction.
DacoClaw is the harness that lets any personal or coding agent — OpenClaw, Hermes, NEMO, Claude Code, Cursor — operate inside a DacoWork workspace through a single MCP gateway.
One set of credentials. One brain. The same governance rules apply to your agent and ours: audit trails, scopes, and the intent guardian are uniform across humans and agents.
We build in the open, dogfooded. What works is discovering what to build and delivering with you in the loop. The fully autonomous factory is the next leap.
Capture, triage and the intelligence layer over your channels. Value delivered with a human in the loop.
The intelligence repository, the intent guardian and the workspace voice already run in code, with MCP and white-label.
The first gap is closed: the intelligence already detects your work session and pushes what's relevant, on its own. The full edge — the next step ready, you just approve — is under construction.
Make software construction actually deliver for the first time — closed verification loop, continuity, self-correction.
We're dogfooding this in front of you. The brain you'll use is the brain we use. Early access — by invitation, then in waves.