[Integrations](https://spiral.ai/integrations) › [goose](https://spiral.ai/integrations/goose) › [FLUX.1 SRPO Dev](https://spiral.ai/integrations/goose/flux-1-srpo-dev)

# How to Use FLUX.1 SRPO Dev in goose

FLUX.1 SRPO Dev turns text descriptions into images for portrait concepts, editorial illustrations, and scene studies. It fine-tunes FLUX.1 Dev to improve realism and visual quality, making it useful when lighting, texture, and composition drive the brief.

|  |  |
| --- | --- |
| Model | FLUX.1 SRPO Dev |
| Available through | fal |
| Category | Image |
| Price | $0.0325 per megapixel. `image_size` presets count as 1 MP; custom sizes round up to whole megapixels. Billed on `num_images`, default 1. |
| Returns | image |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with FLUX.1 SRPO Dev in goose

- Read the homepage copy and design tokens, then use FLUX.1 SRPO Dev to create a 1024 × 1024 workshop scene for the page: a potter shaping a clay bowl beside a window, soft side light, visible clay texture, and space above the subject for a heading.
- Use FLUX.1 SRPO Dev to generate a 1024 × 1024 editorial image for the gardening article in this project. Show a person tending seedlings on an apartment balcony, morning sunlight crossing terracotta pots, realistic leaf texture, and a composition that keeps the main action near the center.
- Read the game's environment brief and use FLUX.1 SRPO Dev to create a 1024 × 1024 concept image of its harbor district. Translate the written architecture and weather into a street-level scene with wet stone, moored boats, overcast light, and a clear foreground.
- Use FLUX.1 SRPO Dev to create a 1024 × 1024 still life for the bookstore page: three closed clothbound books beside a ceramic cup on a walnut table, soft window light, visible fabric texture, and generous surrounding space. Add the page title separately in HTML.
- Read the visual brief and use FLUX.1 SRPO Dev to explore its portrait direction at 1024 × 1024. Depict a fictional bicycle mechanic in a workshop, waist-up framing, natural skin texture, worn cotton clothing, and soft window light; try guidance 3.5 and 50 steps.

## How to set up FLUX.1 SRPO Dev in goose

Pick the way that suits you. Each takes about two minutes.

### Use Spiral's MCP server

1. **Add the extension**: In goose Desktop, open the sidebar, click Extensions, then Add custom extension. Enter:

2. **Sign in**: When goose first connects to Spiral, your browser opens. Sign in to Spiral, pick the organization and click Authorize. No API key. If the browser does not open, goose prints the link to open yourself.

### Use the Spiral skill

1. **Send goose this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run FLUX.1 SRPO Dev**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "fal",
    "model": "fal-ai/flux/srpo",
    "input": {
      "prompt": "Mountain guide, sturdy build, wilderness wisdom, alert gaze, technical outdoor gear with rope coils, snow-capped peaks background, crisp mountain lighting,…"
    }
  }'
```

Then paste one of the prompts above: goose finds FLUX.1 SRPO Dev on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/goose

## FLUX.1 SRPO Dev tools and prices

- [FLUX.1 SRPO \[dev\] · Text to Image](https://spiral.ai/tools/fal/fal-ai/flux/srpo) (fal): $0.0325 per megapixel. `image_size` presets count as 1 MP; custom sizes round up to whole megapixels. Billed on `num_images`, default 1.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| image_size |  | landscape_4_3 | The size of the generated image. |
| num_inference_steps | 1 to 50 | 28 | The number of inference steps to perform. |
| seed |  | — | The same seed and the same prompt given to the same version of the model will output the same image every time. |
| guidance_scale | 1 to 20 | 4.5 | The CFG (Classifier Free Guidance) scale is a measure of how close you want the model to stick to your prompt when looking for a related image to show you. |
| num_images | 1 to 4 | 1 | The number of images to generate. |
| output_format | jpeg, png | jpeg | The format of the generated image. |

## What can you do with FLUX.1 SRPO Dev in goose?

Create original scene assets from the project's written design requirements with FLUX.1 SRPO. A coding agent can read the relevant files, draft image briefs, and place selected results in the interface.

### Editorial lead images

Illustrate an article's theme with an invented scene. Specify a clear subject, directional light, and room for a headline.

### Website scene assets

Create original workshop, landscape, or interior imagery from a written design brief. Match the scene's palette and composition to the page.

### Portrait concept boards

Explore fictional portrait subjects with different wardrobe, framing, and lighting treatments before planning a shoot.

### Still-life art direction

Develop concepts around ceramics, flowers, books, or fabric. Describe the materials and light to explore a visual treatment before production.

### Environment concept studies

Turn written descriptions of fictional streets, buildings, and landscapes into scene concepts for a game or story pitch.

### Lighting studies

Compare separate interpretations of a written scene under window light, overcast daylight, or a studio softbox. Use the results to discuss an art direction.

## Overview

- [fal](https://spiral.ai/providers/fal)
- [All goose guides](https://spiral.ai/integrations/goose)

## FLUX.1 SRPO Dev in other agents

- [Amp](https://spiral.ai/integrations/amp/flux-1-srpo-dev)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/flux-1-srpo-dev)
- [Claude](https://spiral.ai/integrations/claude-desktop/flux-1-srpo-dev)
- [Claude Code](https://spiral.ai/integrations/claude-code/flux-1-srpo-dev)
- [Cline](https://spiral.ai/integrations/cline/flux-1-srpo-dev)
- [Codex](https://spiral.ai/integrations/codex/flux-1-srpo-dev)
- [Cursor](https://spiral.ai/integrations/cursor/flux-1-srpo-dev)
- [Devin](https://spiral.ai/integrations/devin/flux-1-srpo-dev)
- [dots](https://spiral.ai/integrations/dots/flux-1-srpo-dev)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/flux-1-srpo-dev)
- [Gemini](https://spiral.ai/integrations/gemini/flux-1-srpo-dev)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/flux-1-srpo-dev)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/flux-1-srpo-dev)
- [Grok](https://spiral.ai/integrations/grok/flux-1-srpo-dev)
- [Grok Build](https://spiral.ai/integrations/grok-build/flux-1-srpo-dev)
- [Hermes](https://spiral.ai/integrations/hermes/flux-1-srpo-dev)
- [Instinct](https://spiral.ai/integrations/instinct/flux-1-srpo-dev)
- [Manus](https://spiral.ai/integrations/manus/flux-1-srpo-dev)
- [Muse](https://spiral.ai/integrations/muse/flux-1-srpo-dev)
- [Muse Code](https://spiral.ai/integrations/muse-code/flux-1-srpo-dev)
- [n8n](https://spiral.ai/integrations/n8n/flux-1-srpo-dev)
- [Notion](https://spiral.ai/integrations/notion/flux-1-srpo-dev)
- [OpenClaw](https://spiral.ai/integrations/openclaw/flux-1-srpo-dev)
- [opencode](https://spiral.ai/integrations/opencode/flux-1-srpo-dev)
- [Perplexity](https://spiral.ai/integrations/perplexity/flux-1-srpo-dev)
- [Raycast](https://spiral.ai/integrations/raycast/flux-1-srpo-dev)
- [VS Code](https://spiral.ai/integrations/vscode/flux-1-srpo-dev)
- [Windsurf](https://spiral.ai/integrations/windsurf/flux-1-srpo-dev)
- [Zapier](https://spiral.ai/integrations/zapier/flux-1-srpo-dev)

## More goose guides

- [FLUX.1 SRPO](https://spiral.ai/integrations/goose/flux-1-srpo)
- [FLUX.1 Schnell](https://spiral.ai/integrations/goose/flux-1-schnell)
- [FLUX.1 Redux \[dev\]](https://spiral.ai/integrations/goose/flux-1-redux-dev)
- [FLUX.1 Schnell Redux](https://spiral.ai/integrations/goose/flux-1-schnell-redux)
- [FLUX.1 Krea Redux](https://spiral.ai/integrations/goose/flux-1-krea-redux)
- [FLUX.2](https://spiral.ai/integrations/goose/flux-2)

## Questions about FLUX.1 SRPO Dev in goose

**Can goose use FLUX.1 SRPO Dev?**

Yes. Add Spiral to goose as an MCP server, and goose can find and run FLUX.1 SRPO Dev like any built-in tool.

**How much does FLUX.1 SRPO Dev cost in goose?**

$0.0325 per megapixel. `image_size` presets count as 1 MP; custom sizes round up to whole megapixels. Billed on `num_images`, default 1. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your goose plan is separate.

**Do I need a fal account to use FLUX.1 SRPO Dev in goose?**

No. Spiral runs FLUX.1 SRPO Dev on its own fal account. You sign in to Spiral once and never paste a key.

**How long does it take to set up FLUX.1 SRPO Dev in goose?**

About two minutes: add the extension, then sign in.

**Does goose wait while FLUX.1 SRPO Dev runs?**

FLUX.1 SRPO Dev runs in the background. goose starts the run, checks on it and gives you the image when it's ready.

**What is FLUX.1 SRPO good at?**

Its main focus is realism and visual quality. Consider it for portrait concepts, editorial imagery, and scene studies where lighting, texture, and composition matter.

**How does FLUX.1 SRPO differ from FLUX.1 Dev?**

SRPO fine-tunes FLUX.1 Dev using human preference signals. Tencent reports improved realism and aesthetic quality in its human evaluation.

**What image size should I start with?**

Start with 1024 × 1024 for a square composition. That is the size used in Tencent's SRPO inference example.

**What guidance and step count should I try?**

Tencent's reference example uses guidance 3.5 and 50 steps. The paper specifically uses guidance 3.5 to avoid oversmoothing and artifacts from high guidance, so increasing guidance is not automatically an improvement.

**How should I write an effective prompt?**

Write a concrete scene in descriptive prose. Include the subject and action, visual medium, setting, lighting, and framing, following Black Forest Labs' guidance for FLUX.

