[Integrations](https://spiral.ai/integrations) › [opencode](https://spiral.ai/integrations/opencode) › [FLUX.2 Klein 4B Base LoRA](https://spiral.ai/integrations/opencode/flux-2-klein-4b-base-lora)

# How to Use FLUX.2 Klein 4B Base LoRA in opencode

FLUX.2 Klein 4B Base LoRA generates images with custom style adaptation. Unlike FLUX.2 Klein 4B Base, it can apply trained LoRA adapters for a learned aesthetic or character.

|  |  |
| --- | --- |
| Model | FLUX.2 Klein 4B Base LoRA |
| Available through | fal |
| Category | Image |
| Price | $0.016432 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.2 Klein 4B Base LoRA in opencode

- Generate a product photo with FLUX.2 Klein 4B Base LoRA: a matte black water bottle on wet slate, soft morning light, 1:1. Tell me the price first. Save the result in ./assets.
- Make four logo concepts for a coffee brand called "Northline" with FLUX.2 Klein 4B Base LoRA.
- Create a 16:9 hero image of a quiet mountain cabin at dusk with FLUX.2 Klein 4B Base LoRA.
- Generate a 9:16 poster with FLUX.2 Klein 4B Base LoRA for a jazz night called "Blue Hour", deep navy and brass tones.
- Generate a flat illustration of a team planning session with FLUX.2 Klein 4B Base LoRA for our blog header, 16:9.

## How to set up FLUX.2 Klein 4B Base LoRA in opencode

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

### Use Spiral's MCP server

1. **Add the server**: Leave out `--global` to add Spiral to the current project only. No header and no client ID or secret: opencode registers itself with Spiral.

```bash
opencode mcp add spiral --global --url https://api.spiral.ai/mcp
```

2. **Sign in**: Your browser opens. Sign in to Spiral, pick the organization and click Authorize, then return to the terminal. No API key. Inside opencode, `/mcps` starts the same sign-in.

```bash
opencode mcp auth spiral
```

3. **Check it**: Spiral shows as connected.

```bash
opencode mcp list
```

### Use the Spiral skill

1. **Send opencode 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.2 Klein 4B Base LoRA**: 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-2/klein/4b/base/lora",
    "input": {
      "prompt": "A serene Japanese garden with cherry blossoms, koi pond, and traditional wooden bridge at golden hour"
    }
  }'
```

Then paste one of the prompts above: opencode finds FLUX.2 Klein 4B Base LoRA on Spiral, shows the price and runs it.

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

## FLUX.2 Klein 4B Base LoRA tools and prices

- [FLUX.2 \[klein\] 4B Base LoRA · Text to Image](https://spiral.ai/tools/fal/fal-ai/flux-2/klein/4b/base/lora) (fal): $0.016432 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 |
| --- | --- | --- | --- |
| negative_prompt |  |  | Negative prompt for classifier-free guidance. Describes what to avoid in the image. |
| guidance_scale | 0 to 20 | 5 | Guidance scale for classifier-free guidance. |
| seed |  | — | The seed to use for the generation. If not provided, a random seed will be used. |
| num_inference_steps | 4 to 50 | 28 | The number of inference steps to perform. |
| image_size |  | landscape_4_3 | The size of the image to generate. |
| num_images | 1 to 4 | 1 | The number of images to generate. |

## What can you do with FLUX.2 Klein 4B Base LoRA in opencode?

Create image concepts that match a project's written design brief and page layout. A coding agent can turn those requirements into detailed scene prompts and place selected outputs into the project for review.

### Website hero image concepts

Generate scenes from a written creative brief, with the subject placed to leave room for a headline. Compare lighting and framing before choosing a direction.

### Product photography concepts

Describe a proposed product, surface, camera angle, and lighting to explore a shoot concept. Use the results to plan photography and art direction.

### Article illustrations

Turn an article's central idea into a visual metaphor, such as a crowded desk for an essay about distraction. Specify the medium and composition in the brief.

### Campaign art direction

Explore the same campaign idea through different settings, lighting treatments, and visual styles. Compare the images before committing to production.

### Fictional environments

Describe a game location or story setting, including architecture, weather, viewpoint, and atmosphere. Generate concept images for a creative review.

### Poster lettering studies

Explore a poster with a short headline and a clear visual subject. Check the generated spelling and redraw the final lettering when precision matters.

### Artwork in a trained style

Apply a trained style LoRA to new subjects for an illustrated campaign or publication. Include the adapter's documented trigger phrase in each scene description.

### Recurring character concepts

Use a trained character LoRA to depict a character in different scenes and poses. Review each result for consistency before assembling a sequence.

## Overview

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

## FLUX.2 Klein 4B Base LoRA in other agents

- [Amp](https://spiral.ai/integrations/amp/flux-2-klein-4b-base-lora)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/flux-2-klein-4b-base-lora)
- [Claude](https://spiral.ai/integrations/claude-desktop/flux-2-klein-4b-base-lora)
- [Claude Code](https://spiral.ai/integrations/claude-code/flux-2-klein-4b-base-lora)
- [Cline](https://spiral.ai/integrations/cline/flux-2-klein-4b-base-lora)
- [Codex](https://spiral.ai/integrations/codex/flux-2-klein-4b-base-lora)
- [Cursor](https://spiral.ai/integrations/cursor/flux-2-klein-4b-base-lora)
- [Devin](https://spiral.ai/integrations/devin/flux-2-klein-4b-base-lora)
- [dots](https://spiral.ai/integrations/dots/flux-2-klein-4b-base-lora)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/flux-2-klein-4b-base-lora)
- [Gemini](https://spiral.ai/integrations/gemini/flux-2-klein-4b-base-lora)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/flux-2-klein-4b-base-lora)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/flux-2-klein-4b-base-lora)
- [goose](https://spiral.ai/integrations/goose/flux-2-klein-4b-base-lora)
- [Grok](https://spiral.ai/integrations/grok/flux-2-klein-4b-base-lora)
- [Grok Build](https://spiral.ai/integrations/grok-build/flux-2-klein-4b-base-lora)
- [Hermes](https://spiral.ai/integrations/hermes/flux-2-klein-4b-base-lora)
- [Instinct](https://spiral.ai/integrations/instinct/flux-2-klein-4b-base-lora)
- [Manus](https://spiral.ai/integrations/manus/flux-2-klein-4b-base-lora)
- [Muse](https://spiral.ai/integrations/muse/flux-2-klein-4b-base-lora)
- [Muse Code](https://spiral.ai/integrations/muse-code/flux-2-klein-4b-base-lora)
- [n8n](https://spiral.ai/integrations/n8n/flux-2-klein-4b-base-lora)
- [Notion](https://spiral.ai/integrations/notion/flux-2-klein-4b-base-lora)
- [OpenClaw](https://spiral.ai/integrations/openclaw/flux-2-klein-4b-base-lora)
- [Perplexity](https://spiral.ai/integrations/perplexity/flux-2-klein-4b-base-lora)
- [Raycast](https://spiral.ai/integrations/raycast/flux-2-klein-4b-base-lora)
- [VS Code](https://spiral.ai/integrations/vscode/flux-2-klein-4b-base-lora)
- [Windsurf](https://spiral.ai/integrations/windsurf/flux-2-klein-4b-base-lora)
- [Zapier](https://spiral.ai/integrations/zapier/flux-2-klein-4b-base-lora)

## More opencode guides

- [FLUX.2 Klein 4B Base](https://spiral.ai/integrations/opencode/flux-2-klein-4b-base)
- [FLUX.2 Klein 4B LoRA](https://spiral.ai/integrations/opencode/flux-2-klein-4b-lora)
- [FLUX.2 Klein 4B](https://spiral.ai/integrations/opencode/flux-2-klein-4b)
- [FLUX.2 Klein 9B](https://spiral.ai/integrations/opencode/flux-2-klein-9b)
- [FLUX.2 HDR Style](https://spiral.ai/integrations/opencode/flux-2-hdr-style)
- [FLUX.2 Klein 9B Base](https://spiral.ai/integrations/opencode/flux-2-klein-9b-base)

## Questions about FLUX.2 Klein 4B Base LoRA in opencode

**Can opencode use FLUX.2 Klein 4B Base LoRA?**

Yes. Add Spiral to opencode as an MCP server, and opencode can find and run FLUX.2 Klein 4B Base LoRA like any built-in tool.

**How much does FLUX.2 Klein 4B Base LoRA cost in opencode?**

$0.016432 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 opencode plan is separate.

**Do I need a fal account to use FLUX.2 Klein 4B Base LoRA in opencode?**

No. Spiral runs FLUX.2 Klein 4B Base LoRA 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.2 Klein 4B Base LoRA in opencode?**

About two minutes: add the server, then sign in, then check it.

**Does opencode wait while FLUX.2 Klein 4B Base LoRA runs?**

FLUX.2 Klein 4B Base LoRA runs in the background. opencode starts the run, checks on it and gives you the image when it's ready.

**What is FLUX.2 Klein 4B Base good at?**

It is useful for realistic image concepts and exploring different visual treatments of a written brief. Its undistilled design gives it more output diversity than distilled Klein models.

**How does Base differ from distilled Klein?**

Base preserves the full training signal and favors flexibility and control. Treat it as a model for deliberate visual exploration; the maker's 4B Base example uses 50 inference steps.

**How should I write a prompt?**

Describe a scene in connected sentences, including the subject, setting, lighting, composition, and style. Black Forest Labs recommends narrative descriptions instead of keyword lists.

**What image size is a documented starting point?**

The maker demonstrates 1024 by 1024 output for 4B Base. Use that square canvas for an initial comparison, then compose for the shape your project needs.

**Can I trust generated lettering?**

Review every word before using the image. The model can render text, but its model card warns that lettering can be inaccurate or distorted.

