> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.withpaygent.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.withpaygent.com/_mcp/server.

# Custom Pricing

Paygent ships with built-in cost tables for popular LLM, STT, TTS, and video models. **Custom pricing** lets you override those defaults — or define entirely new provider/plan/model combinations — so revenue calculations match your commercial rates.

Use custom pricing when you:

* Resell models at a markup above provider cost
* Negotiate enterprise rates per customer (via [Plans](/documentation/introduction/plans))
* Bill voice agents per minute with custom LLM/STT/TTS revenue rates
* Price video generation by resolution tier

---

## Two layers of custom pricing

| Layer                          | Where                         | Scope                                             |
| ------------------------------ | ----------------------------- | ------------------------------------------------- |
| **Organization custom models** | Settings → **Custom Pricing** | Global rate cards by provider / plan / model      |
| **Indicator custom pricing**   | Agent → indicator pricing     | Per-agent voice rates + linked configuration name |

Both layers work together: organization configs define model token/minute rates; indicator configs define per-minute revenue for voice stacks.

---

## Organization custom model pricing

Sidebar → **Custom Pricing**

Create rate cards keyed by:

| Dimension      | Examples                                         |
| -------------- | ------------------------------------------------ |
| **Model type** | LLM, STT, TTS, Video                             |
| **Provider**   | OpenAI, Anthropic, ElevenLabs, Google            |
| **Plan**       | Provider tier (e.g. `standard`, `batch`)         |
| **Model**      | `gpt-4o`, `claude-3-5-sonnet`, `eleven_turbo_v2` |

### LLM rates (per 1K tokens)

| Field         | Description                           |
| ------------- | ------------------------------------- |
| Input tokens  | Revenue/cost per 1K input tokens      |
| Output tokens | Per 1K output tokens                  |
| Cached tokens | Per 1K cached tokens (when supported) |

### STT / TTS rates

| Field                  | Description                 |
| ---------------------- | --------------------------- |
| Audio per minute       | Rate per minute of audio    |
| Characters per million | TTS character-based pricing |

### Video rates

| Field              | Description                                         |
| ------------------ | --------------------------------------------------- |
| Default resolution | Fallback when resolution is unknown                 |
| Resolution tiers   | Per-resolution rates for video and video-with-audio |

Add resolution tiers like `720x1280`, `1920x1080`, or `4k` with separate video / video-with-audio prices.

### Bulk import

Use the **Bulk JSON** tab to import many configs at once. Each entry needs `modelType`, `provider`, `plan`, `model`, and the rate fields for that type.

> **Info**
>
> Custom model configs are organization-wide. They apply whenever Paygent resolves cost/revenue for matching provider + model + plan on tracked usage events.

---

## Indicator-level custom pricing (voice)

For voice and speech-to-speech agents, configure **custom pricing on the indicator** under agent pricing:

| Field                   | Description                                                 |
| ----------------------- | ----------------------------------------------------------- |
| **Configuration name**  | Links to an organization custom model config (optional)     |
| **Default rates**       | Fallback LLM / STT / TTS revenue per minute                 |
| **Custom pricing data** | Per-provider or per-model overrides with `revenuePerMinute` |

This is how you charge customers a flat **\$0.15/min** for a voice call while Paygent still tracks underlying provider costs separately for margin analysis.

---

## How custom pricing affects billing

```
SDK sends usage (tokens, minutes, video metadata)
        │
        ▼
Paygent resolves rates:
  1. Organization custom model config (if match)
  2. Built-in provider defaults (fallback)
  3. Indicator custom pricing (voice revenue override)
        │
        ▼
Revenue calculated → invoice line items / analytics
```

Custom pricing affects **revenue** (what you charge). Provider costs still flow to cost dashboards for unit economics unless you only track revenue-only indicators.

---

## Example: markup on GPT-4o

1. **Custom Pricing** → Model type LLM → Provider OpenAI → Plan standard → Model `gpt-4o`
2. Set input `$0.005`/1K, output `$0.020`/1K (your sell rates)
3. Agent indicator uses automatic LLM tracking
4. Customer usage bills at your rates; dashboard shows margin vs provider cost

---

## Example: voice agent per-minute revenue

1. Create voice agent with indicator `call-completed`
2. On indicator pricing, open **Custom pricing**
3. Set default LLM revenue `$0.08/min`, STT `$0.02/min`, TTS `$0.05/min`
4. Or link a configuration name from organization custom models
5. SDK tracks voice session → revenue summed per minute on invoice

See [Voice Tracking](/documentation/voice-tracking/voice-agent-setup) for integration details.

---

## Validation and errors

Paygent validates custom pricing on save:

* Required fields must be filled for the selected model type
* Video configs require at least one resolution tier
* Pricing values cannot be negative

If a usage event references a model with no matching config, Paygent falls back to built-in defaults — usage is still tracked; revenue may differ from your intended rate until you add the config.

---

## Checklist

* ✓ Organization custom model configs created for models you resell
* ✓ Voice agents: indicator custom pricing or default per-minute rates set
* ✓ Agent/plan indicator billing type matches your commercial model (FLAT, CREDITS, etc.)
* ✓ SDK sends usage with correct `serviceProvider` and `model` for rate matching

**Related:** [Pricing Models](/documentation/introduction/pricing-models) · [Voice Tracking](/documentation/voice-tracking/voice-agent-setup) · [Cost & Revenue Tracking](/documentation/guides/cost-revenue-tracking)