---
title: "Use Seedream in Hermes Agent via RunAPI — Image API Guide"
url: "https://runapi.ai/hermes-seedream.md"
canonical: "https://runapi.ai/hermes-seedream"
locale: "en"
model: "seedream"
---

# Use Seedream in Hermes Agent.

Seedream is Bytedance's image generation model with strong typography rendering and up to 4K output. Hermes Agent calls it through the same RunAPI custom provider and API key used for chat — no extra plugins or GPU rentals.

## API example

```bash
curl -X POST https://runapi.ai/api/v1/task/text_to_image \
  -H "Authorization: Bearer $RUNAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedream-4.5-text-to-image",
    "prompt": "A premium coffee bag on a marble countertop, warm morning light, bold serif text FRESH ROAST, 4K product photography"
  }'

```

### Response

```json
{
  "task_id": "tsk_abc123",
  "status": "pending",
  "model": "seedream-4.5-text-to-image"
}

```

## How it works

1. **Configure RunAPI** — If you already set up RunAPI in Hermes Agent for chat, you are done. The same custom:runapi provider and RUNAPI_API_KEY handle image generation. If not, add the custom provider with base_url https://runapi.ai/v1 and run hermes doctor.
2. **Call Seedream** — Send a request to the text_to_image endpoint with model set to seedream-4.5-text-to-image. Include a text prompt describing the image. RunAPI returns a task ID immediately.
3. **Get the result** — Poll the task status endpoint with the returned task ID. When the status changes to completed, the response includes the generated image URL. RunAPI SDKs handle polling automatically.

## Parameters

| Parameter | Type | Description |
|-----------|------|-------------|
| `model` | `string` | Required. seedream-4.5-text-to-image, seedream-4.5-edit, seedream-5-lite-text-to-image. |
| `prompt` | `string` | Text description of the desired image. |
| `size` | `string` | Optional. Output dimensions. Supports up to 4K. |
| `image_url` | `string` | Source image for editing mode. |

## FAQ

### Can I use Seedream in Hermes Agent?

Yes. Configure RunAPI as a custom provider in Hermes Agent with base_url https://runapi.ai/v1, then call the text_to_image endpoint with model set to seedream-4.5-text-to-image. The same API key handles both chat and image generation.

### Do I need a separate API key for Seedream?

No. The same RUNAPI_API_KEY you configured for the custom:runapi provider also calls Seedream and every other image model in the RunAPI catalog. One key, one billing account.

### How good is the text rendering compared to other image models?

Seedream achieves roughly 94% text accuracy, which puts it among the top tier for typography in generated images. It handles multi-line text, mixed fonts, and text on curved surfaces better than most competitors. For the absolute best text rendering, compare with Ideogram V3 on the same prompt.

### Can Hermes Agent use Seedream in a multi-step design pipeline?

Yes. Hermes Agent can orchestrate Seedream for initial generation, pass the output to Flux Kontext for targeted edits, and then route the result to Topaz for final upscaling, all in a single agent workflow.

### Can Seedream maintain character consistency across a batch?

Yes. Describe the character in detail in each prompt and Seedream maintains facial features, proportions, and styling across outputs. This makes it practical for social media campaigns where a brand ambassador needs to appear consistently across dozens of images.

### What makes Seedream different from Imagen 4 or Nano Banana?

Seedream is strongest at maintaining a consistent house style across large batches and handles complex multi-object prompts well. Imagen 4 offers better value per image at the low end with its Fast tier. Nano Banana leads on world knowledge integration. Pick based on whether you need batch consistency, low cost, or contextual awareness.


## Links

- [Hermes Agent setup guide →](https://runapi.ai/hermes-agent)
- [Seedream models →](https://runapi.ai/models/seedream)
- [Model catalog](https://runapi.ai/models)
- [API docs](https://runapi.ai/docs)