> ## Documentation Index
> Fetch the complete documentation index at: https://getkontext.io/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Usage & overages

> What counts, the caps, and how overages behave.

## Feedback items

**What counts.** A feedback item is counted **every time it is processed**, whether it's a brand-new submission or an update to an existing one. Feedback evolves: a support ticket gets new replies, a Linear issue gets new comments, a transcript gets appended to. Each of those kicks off a fresh analysis run and adds one to the counter.

**Where it's counted.** Ingestion happens from the UI, the [public API](/docs/api-reference/overview), the [MCP server](/docs/mcp/overview), and every connected [integration](/docs/integrations/overview). All paths feed the same counter. If you submit the same feedback twice, it counts twice. Items that fail analysis aren't refunded.

### Included amounts and the hard cap

Each plan ships an **`included`** amount per period. On top of that sits a **hard cap at twice the included amount**, a safety lid that can't be crossed even with overages enabled. It exists to protect you from runaway billing (a misconfigured webhook, an accidental re-ingest loop). If you genuinely need it lifted, [contact us](https://getkontext.io/contact).

| Plan       | Included / period | Hard cap |
| ---------- | ----------------- | -------- |
| Starter    | 1,000             | 2,000    |
| Growth     | 10,000            | 20,000   |
| Business   | 25,000            | 50,000   |
| Enterprise | Unlimited         | —        |

### Overages, opt-in and off by default

Toggle overages from **Organization Settings → Billing → Overages**.

**Overages off (the default).** Reaching the `included` cap rejects new ingestion with an upgrade-or-enable-overages error. You stop ingesting rather than quietly run up a tab.

**Overages on.** Ingestion keeps running past `included` up to the hard cap. Each item above `included` is billed on your next invoice.

| Plan       | Feedback overage per item | Seat overage per seat / month |
| ---------- | ------------------------- | ----------------------------- |
| Starter    | Not available             | \$7                           |
| Growth     | \$0.02                    | \$7                           |
| Business   | \$0.02                    | Unlimited (no overage)        |
| Enterprise | Negotiated                | Negotiated                    |

**No feedback overage on Starter.** If you consistently exceed 1,000 items per period, the right move is to upgrade to Growth rather than lean on overages.

## Seats

**What's a seat.** One seat covers one organization member. Active members count, and **pending invitations count too** (an outstanding invite reserves a seat until it's revoked or the member joins). Deactivated members don't count.

**Adding seats above included.** You're charged immediately, prorated to the end of the current billing period.

**Removing an over-limit seat.** The unused portion is credited on your next invoice.

**Business and Enterprise.** Both include unlimited seats, so there's no seat overage because there's no cap.

## The preconfigured LLM allowance

Kontext runs on large language models. In normal operation you **bring your own key** (BYOK): configure a provider (OpenAI, Anthropic, your own gateway) in **Organization Settings → AI Providers**, and all analysis runs against it on your bill.

**Why the allowance exists.** To avoid forcing BYOK configuration before you've seen Kontext work, every plan ships a small preconfigured allowance of LLM calls made against a provider that Kontext owns and pays for. It's enough to ingest a handful of feedback items and see the full analysis pipeline in action.

**Hard-capped, by design.** Once the allowance is exhausted, further analysis requires a configured BYOK provider. There's no on-by-default overage path on the preconfigured LLM. The goal is to lower the bar to the first productive use of Kontext, not to be a permanent hosted-LLM service.

**Need more?** If you have a genuine reason to keep running on the preconfigured LLM past the allowance (evaluation stretch, BYOK not yet approved internally), [contact us](https://getkontext.io/contact) and we'll work something out. Just don't plan on it as a long-term setup.

## Monitoring your usage

**In the app.** Current-period counters, with the included line and hard-cap line drawn, live under **Organization Settings → Usage**.

**Upcoming charges.** With overages enabled, the accumulating overage charge for the current period is visible on your upcoming invoice in the billing portal, so there's no guessing about where you'll land.

## Related

* [Plan changes](/docs/billing/plan-changes) explains what happens to usage when you switch plans mid-cycle, or downgrade while already over the new plan's cap.
* [Trial](/docs/billing/trial) uses the same cap behavior but overages don't apply.