Your plan isn't just a price and an AI token quota (you already know that from **Manual 34 · Billing**): it also sets **how many of each kind of thing** you can have active at once, and which parts of the product are included. Each of those things is a **capability**, and it's always one of three types: | Type | What it measures | Examples | | --- | --- | --- | | **Yes/No** (on or off) | Whether that part of the product is included in your plan. | Public API; each specific integration (HubSpot, Google Sheets…). | | **Quantity** (live instances) | How many you can have created RIGHT NOW — not how many you've ever created. | AI Agents, Flows, Tasks, channels by type (WhatsApp, Telegram, Web Widget…), API keys, MCP servers, document templates. | | **Quota** (period consumption) | How much you can use each month — it resets with your billing cycle. | Conversations this month, marketing sends this month. | > **Note — The core of the product never has a volume limit** > > Leads, contacts, calendar, tasks, analytics, knowledge base, documents, support and settings work the same in every plan, with no cap on “how many leads you can have” or “how many appointments you can book” — that part of the product isn't a limitable capability. The only thing in the core with a real quota is **consumption**: conversations and marketing sends for the month (section 2), because that does have a real cost behind it. Every **Quantity** capability counts **live instances**, never a cumulative history: if your plan includes 3 agents and you delete one of the 3 you had, you can create another one again — the counter goes down with the deleted object, not just up when you create one.