Adros

The marketing operating system that turns real ad accounts and strategy into traceable work.

Adros has three layers:

1. Deterministic rails - tasks, event timelines, approvals, blockers, client-scoped files, and scoped integrations. Work is auditable whether it starts in the dashboard or via MCP.
2. Managed-agent depth - when the problem is ambiguous, Adros routes work to the right specialist instead of pretending every problem is a one-shot tool call.
3. Operator surfaces - a full dashboard at app.adros.ai (opens in a new tab) where humans see the same spine the agents use: Command Center, Strategy Workspace, Files, Campaigns, and Adros CEO.

What's inside

Operator surfaces (dashboard)

Backend capabilities

• 4,022 performance-tested ad patterns - creative moat across styles and frameworks
• Managed Agents - adros_research (keyword, market, competitor, and audience research), adros_cmo (audits, strategy synthesis, structure, optimization), adros_cd (copy/creative), adros_deployer (launch/changes), adros_ceo (routing/intake)
• Daily autonomous monitor - 6 health checks with HMAC-signed webhook delivery
• Quality track - audit scorecards, guardrails, and artifact persistence on completed audits
• Contract layer - deterministic deliverable specs per lifecycle stage
• Creative pipeline - generation anchored to pattern blueprints and product guardrails
• Memory layer - business context, personas, weekly logs scoped per client
• ~74 MCP tools - Streamable HTTP at /mcp for Claude, Cursor, or custom agents
• Artifact canon - versioned analysis, mastered operating specs, traceable from strategy through deploy

Production vs repo

Dashboard and backend features ship continuously. What you see on app.adros.ai / api.adros.ai may lag the latest Git main by one deploy. For integration testing, verify against your target environment and trust the live API if this site and the deployment diverge.

Get started

MCP (3 minutes)

{
  "mcpServers": {
    "adros": {
      "url": "https://api.adros.ai/mcp",
      "headers": { "X-User-Token": "lyr_YOUR_TOKEN" }
    }
  }
}

1. Sign up at app.adros.ai (opens in a new tab) and copy your MCP token from Settings -> MCP
2. Paste the JSON above into your AI client's MCP config
3. Restart the client - Adros tools load automatically

Dashboard (no setup)

Sign in at app.adros.ai (opens in a new tab) and use Command Center, Strategy Workspace, Files, Campaigns, and CEO directly. Same backend, same auth model, same tasks and artifacts.

-> Full step-by-step: Quickstart

Where to go from here

Connection details (reference)

Who makes this

Adros is built by Matrix AI Solution Pte Ltd (opens in a new tab) in Singapore. If anything on this site contradicts the deployment you are using, trust the deployment and file an issue.