---
name: prompt-adapt
description: >
  Adapt and convert AI prompts between different models and platforms.
  Translates between Midjourney, Flux, Leonardo AI, DALL-E, Sora, Imagen,
  Stable Diffusion, Adobe Firefly, Ideogram, and more. Handles syntax
  differences, parameter mapping, and model-specific optimizations.
  Use when user says "adapt prompt", "convert prompt", "translate prompt",
  "port to flux", "midjourney to dall-e", or "change model".
---

# Prompt Adapter

Convert prompts between AI models while preserving intent and maximizing output quality.

## Adaptation Workflow

### Step 1: Identify Source and Target

Determine:
1. **Source model**: What model was this prompt written for?
2. **Target model**: What model should it run on?
3. **Priority**: Preserve style fidelity or optimize for target strengths?

### Step 2: Analyze Source Prompt

Break down the prompt into components:
- Core subject/action
- Style modifiers
- Technical parameters (model-specific)
- Negative prompts (if any)
- Aspect ratio / dimensions

### Step 3: Apply Model Translation Rules

Load `{PROMPT_ENGINE_DIR}/references/model-guide.md` for detailed rules. Key translations:

**Midjourney -> Flux:**
- Remove `--ar`, `--v`, `--style`, `--s`, `--chaos` parameters
- Expand shorthand into natural language descriptions
- Flux prefers longer, more descriptive prompts
- Remove `::` weight syntax, integrate naturally

**Midjourney -> DALL-E:**
- Remove all `--` parameters
- Rewrite as clear, direct descriptions
- DALL-E prefers straightforward language over artistic jargon
- Remove negative prompts (DALL-E doesn't support them well)

**Flux -> Midjourney:**
- Add `--ar` for aspect ratio
- Add `--v 6.1` or appropriate version
- Condense long descriptions into key phrases
- Add style parameters (`--style raw` for photorealistic)

**Any -> Sora (Video):**
- Add camera movement descriptions (pan, zoom, tracking, etc.)
- Add temporal flow ("the scene transitions from... to...")
- Specify duration if possible
- Focus on motion and action over static details

**Any -> Leonardo AI:**
- Reference specific Leonardo models (Phoenix, Alchemy, etc.)
- Use Leonardo-specific quality tokens
- Adapt negative prompts to Leonardo format

### Step 4: Search for Target Model Examples

Find reference prompts in the target model:
```bash
python3 {PROMPT_ENGINE_DIR}/scripts/search_prompts.py "SUBJECT" --model TARGET_MODEL --limit 3
```

Use these as style references for the adaptation.

### Step 5: Present Adaptation

Output format:
1. **Original prompt** (source model labeled)
2. **Adapted prompt** (target model labeled)
3. **Translation notes** (what changed and why)
4. **Parameter mapping** (source params -> target params)
5. **Confidence level** (High/Medium/Low -- based on model compatibility)

## Common Pitfalls

- Midjourney weight syntax (`::2`) has no direct equivalent in most models
- DALL-E ignores most style parameters -- weave them into descriptions
- Sora needs temporal language that image models don't use
- Aspect ratios must be specified differently per platform
- Some styles only work well on specific models (e.g., `--niji` is Midjourney-only)


---
Source attribution: AgriciDaniel
Original project: https://github.com/AgriciDaniel/claude-prompts
License: MIT
License information: https://opensource.org/license/mit
Original upstream instructions; downloaded from Dave AI.

## Upstream license notice

MIT License

Copyright (c) 2026 Gabriel Moreira

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.



# Dave AI resource metadata

Creator / publisher: AgriciDaniel
Original project: https://github.com/AgriciDaniel/claude-prompts
Attribution basis: Upstream repository explicitly recorded in package mirror path

Fields of work: Design, Marketing
Roles: Content Marketer, Graphic Designer, Influencer Marketer, Product Marketer, UX Designer
Specialties: Specialty not specified
Tools: DALL-E, Midjourney, Stable Diffusion
Country applicability: Country not specified
Classification basis: Source catalog labels
Instruction SHA-256: 448a535f26f9764f14ea401d9f2f8eec65549a3de198974c3ddf340dbbcd7db4