# How to Use Plain Text Prompt Templates for Manual Use

> Discover how to use plain text prompt templates from the awesome-gpt-image-2 repository. Copy, customize, and submit prompts easily via cURL or the web UI without coding.

- Repository: [苍何/awesome-gpt-image-2](https://github.com/freestylefly/awesome-gpt-image-2)
- Tags: how-to-guide
- Published: 2026-09-05

---

**The awesome-gpt-image-2 repository stores ready-made plain text prompt templates in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) that you can copy, customize with your own descriptions, and submit directly via cURL or the web UI without writing any application code.**

The freestylefly/awesome-gpt-image-2 project provides a curated collection of **plain text prompt templates** designed specifically for manual use. These templates reside as simple text files within the repository's documentation folder, allowing you to generate images by editing placeholder text and sending the result to the generation endpoint. This approach requires no programming knowledge beyond basic text editing and API calls.

## Locating the Template Files

All **plain text prompt templates** are centralized in **[`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md)** at the repository root. This file organizes templates into distinct sections including *Basic Prompt*, *Style Prompt*, and *Advanced Prompt*, each containing concrete examples you can copy verbatim.

The templates use a simple structure combining descriptive text with optional command-line style modifiers. When you open [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md), you will find text blocks containing placeholders like `<YOUR SUBJECT>` alongside flags such as `--v 5` or `--ar 16:9` that control generation parameters.

## Understanding Template Structure and Modifiers

Before sending a template to the API, you must understand the two core components of each plain text prompt:

- **Descriptive Content**: The natural language description of the image you want to generate, which may include placeholders like `<YOUR SUBJECT>` that you replace with specific content.
- **Generation Modifiers**: Command-line style flags appended to the prompt that control technical parameters. Common modifiers include:
  - `--v 5`: Specifies the model version (version 5)
  - `--ar 16:9`: Sets the aspect ratio to 16:9
  - `--stylize 750`: Adjusts the stylization intensity (higher values = more artistic)
  - `--quality 2`: Specifies rendering quality level

According to the source code in [`api/generate-image.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/api/generate-image.js), the backend parses these modifiers from the plain text string before forwarding the request to the image generation model.

## Step-by-Step Manual Usage

Follow these steps to use the **plain text prompt templates** for manual image generation:

### Step 1: Select and Copy a Template

Open [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) and choose a template matching your desired output style. Copy the entire text block, including any modifiers at the end. For example:

```text
A serene mountain lake at sunset, photorealistic, --v 5 --ar 4:3

```

### Step 2: Customize the Placeholders

Replace any placeholder text such as `<YOUR SUBJECT>` with your specific content. You can also adjust modifiers to change the output characteristics. For instance, changing `--ar 4:3` to `--ar 16:9` switches the image orientation from standard to widescreen.

### Step 3: Send via cURL

Submit the customized prompt to the API endpoint using a cURL command. The server-side handler in [`api/generate-image.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/api/generate-image.js) expects a JSON payload with a `prompt` field:

```bash
curl -X POST https://api.example.com/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt":"A serene mountain lake at sunset, photorealistic, --v 5 --ar 4:3"}'

```

Replace `https://api.example.com/generate` with your actual endpoint URL as configured in the repository's environment.

### Step 4: Send via Web UI

Alternatively, paste the edited **plain text prompt** directly into the "Prompt" field on the generation page of the web interface. Click the *Generate* button to submit the request. The backend processes the raw text identically to the API method, parsing modifiers and generating the image.

## Working Code Examples

Here are three practical examples demonstrating different template categories from [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md):

### Example 1: Basic Prompt

This template uses minimal modifiers for straightforward generation:

```text
A serene mountain lake at sunset, photorealistic, --v 5 --ar 4:3

```

**cURL implementation:**

```bash
curl -X POST https://api.example.com/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt":"A serene mountain lake at sunset, photorealistic, --v 5 --ar 4:3"}'

```

### Example 2: Style Prompt

This example incorporates artistic direction through the `--stylize` parameter:

```text
A portrait of a cyber-punk detective, neon colors, comic-book style, --v 4 --stylize 750

```

**cURL implementation:**

```bash
curl -X POST https://api.example.com/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt":"A portrait of a cyber-punk detective, neon colors, comic-book style, --v 4 --stylize 750"}'

```

### Example 3: Advanced Prompt with Multiple Modifiers

This template combines multiple technical parameters for maximum control:

```text
An otherworldly forest with floating islands, ultra-wide lens, HDR, --v 5 --ar 21:9 --quality 2

```

**cURL implementation:**

```bash
curl -X POST https://api.example.com/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt":"An otherworldly forest with floating islands, ultra-wide lens, HDR, --v 5 --ar 21:9 --quality 2"}'

```

## Local Development and Testing

For local testing before deploying to production, use the **`scripts/vite-local-api.mjs`** script included in the repository. This script launches a local development server that accepts **plain text prompts** identical to the production endpoint. Run the script from your terminal, then direct your cURL commands to `http://localhost:3000/generate` (or the port specified in the script output) to test your customized templates without consuming production API credits.

## Summary

- **Plain text prompt templates** are stored in [`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md) within the freestylefly/awesome-gpt-image-2 repository.
- Templates support placeholders like `<YOUR SUBJECT>` and modifiers including `--v`, `--ar`, `--stylize`, and `--quality`.
- Submit templates manually via cURL by sending JSON to the endpoint handled by [`api/generate-image.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/api/generate-image.js).
- Use the web UI by pasting raw template text into the prompt field for code-free generation.
- Test locally using `scripts/vite-local-api.mjs` before production deployment.

## Frequently Asked Questions

### Where are the plain text prompt templates located in the repository?

All templates are centralized in **[`docs/templates.md`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md)** at the repository root. This file contains sections for Basic, Style, and Advanced prompts that you can copy and customize.

### What do the modifier flags like `--v` and `--ar` control?

The `--v` flag specifies the model version (e.g., `--v 5` for version 5), while `--ar` sets the aspect ratio (e.g., `--ar 16:9` for widescreen). Additional flags like `--stylize` control artistic intensity and `--quality` sets the rendering quality level, all parsed by the backend handler in [`api/generate-image.js`](https://github.com/freestylefly/awesome-gpt-image-2/blob/main/api/generate-image.js).

### Can I use these templates without writing any code?

Yes. You can paste the **plain text prompt** directly into the web UI's prompt field and click *Generate*. Alternatively, use the provided cURL commands by only modifying the text content within the JSON payload, requiring no programming beyond basic text editing.

### How do I test templates locally before using production API credits?

Run **`scripts/vite-local-api.mjs`** to start a local development server. This script creates an endpoint on your local machine that accepts the same plain text prompts as the production API, allowing you to iterate on template customizations without external API calls.