Flottix v2.0: MCP for Everyone — Full Agent Access on Every Plan

Seven months ago, in v1.3, we gave Flottix a second brain: a production-grade MCP server that let AI assistants like Claude read and operate your IT inventory.

Then we made a mistake. We gated it.

Free plans got no API. Starter got read-only. Full read/write lived behind the Professional tier. The people most excited about an MCP-native asset registry — the developers, the solo IT admins, the agent builders — signed up, hit a 403, and left. We built our best feature and locked the door.

Today we’re fixing that.

Full MCP, Every Plan — Including Free

Starting with v2.0, every Flottix plan ships the complete MCP server:

• 17 purpose-built tools covering 70+ operations — one typed, verb-dispatched tool per domain (assets, licenses, kits, peripherals, components, consumables, people, directory, insights, governance), plus compound workflows for onboarding, offboarding, zombie-asset audits, and procurement ingestion
• A REST escape hatch — flottix_api reaches any endpoint, and the whole API is published as a formally valid OpenAPI 3.1 contract your agent can read
• Full read/write — check-out, check-in, assignments, status updates, the whole lifecycle
• Unlimited tool calls — no metering, no throttles beyond sane rate limits
• Free for up to 50 items, forever — no credit card, no trial countdown

If you already run your fleet elsewhere, the Free tier is a real home for a solo IT admin — not a demo.

Why We Removed the Paywall

The economics of MCP-native software are unusual, and they’re the reason this works:

The agent’s intelligence runs on your subscription, not ours. When Claude reasons about your fleet, that compute is yours. Flottix serves deterministic, structured data — which costs us almost nothing for a 50-item workspace. Gating the protocol bought us nothing and cost us everything: the exact users who would evangelize the product couldn’t experience its superpower.

So we changed what we charge for:

We monetize fleet volume and governance. Never the protocol.

Agent-Native ITAM Is a Safety Story

“Let an AI write to my inventory” sounds terrifying — unless the system underneath is deterministic. That’s the entire thesis of Flottix:

1. The LLM proposes. Agents are great at intent: “offboard Paul”, “find waste”, “assign this MacBook”.
2. Flottix disposes. A state machine enforces tenant isolation, license seat limits, and foreign-key validity. An agent cannot invent who possesses a MacBook.
3. You stay accountable. Every action — human or agent — lands in an immutable audit trail. Dangerous operations use two-phase confirmation: the agent presents an impact manifest, waits for your confirmation, then executes.

With Professional, background agents get Staged Approvals: proposals queue up, and an admin reviews the diff before anything executes.

Connect Your Agent in Under Two Minutes

Generate an API key in Settings → API Keys, then point your MCP client at the Flottix endpoint — Cursor, Claude Code, VS Code, Cline, and Gemini CLI all connect natively with one URL and one header:

{
  "flottix": {
    "url": "https://flottix.app/api/mcp",
    "headers": { "Authorization": "Bearer flottix_YOUR_KEY" }
  }
}

Claude Code users:

claude mcp add --transport http flottix https://flottix.app/api/mcp \
  --header "Authorization: Bearer flottix_YOUR_KEY"

Programmatic clients can exchange the key for a short-lived OAuth 2.1 token (one hour, audience-bound, scope-narrowed) — with RFC 9728 discovery for spec-compliant clients.

Then just ask: “Which laptops are assigned to people who left the company?” — and watch your agent work the registry.

👉 Start free with full MCP access

To everyone who told us the paywall was self-sabotage: you were right. Thank you.