> ## Knowledge Base Index > Fetch the complete knowledge base index at: https://help.originality.ai/sitemap.xml > Use this file to discover available pages before exploring further. > Pure-Markdown content can be obtained by appending a '.md' suffix to the content URLs listed in the sitemap (without the trailing slash). # Prompt Generator The **Prompt Generator** is a powerful addition to the Originality.ai platform. It enables users to create optimized prompts for leading large language models (LLMs), fine-tune them for specific providers, and execute them instantly within the app. The tool is designed for speed, usability, and flexibility. * Subscription is NOT required to use the Prompt Generator. * Pay-As-You-Go (PAYG) users can generate, tune, and execute prompts using credits. ${youtube}[Prompt Generator](pON1S957R6M) ### Accessing the Prompt Generator You can access the Prompt Generator from the Settings menu inside the app. ![Accessing the Prompt Generator](https://storage.crisp.chat/users/helpdesk/website/-/b/e/f/3/bef330df68156000/screenshot-2026-02-25-at-10332_14wvo0c.png) ## Core Features ## Prompt Generation Panel 1. User Input Enter a task or idea (3–250 words). 2. Generate Prompt Click **“Generate Prompt”** to create a structured prompt optimized for LLMs. * **Cost:** 1 credit per generation * Prompt generation is powered by Google’s Gemini models. While highly reliable, rare provider outages may temporarily impact generation. ![Prompt Generation Panel](https://storage.crisp.chat/users/helpdesk/website/-/b/e/f/3/bef330df68156000/screenshot-2026-02-25-at-10274_ezoqpx.png) ## Prompt Results Panel Once generated, you can: 1. View the Generated Prompt Your structured prompt appears in the results panel. 2. Copy Prompt One-click copy to clipboard. 3. Recommended Models The system automatically classifies your prompt as: * Text output * Image output Based on this classification, two compatible models are recommended. 4. Fine Tuning Fine tuning restructures the prompt to match the preferred syntax and format of the selected model provider. * Uses provider-specific formatting best practices * Can be undone at any time **Cost:** 1 credit per tune 5. Prompt Execution Click **“Execute Prompt”** to run the displayed prompt on the selected model. * If fine-tuned, the tuned version executes * If reverted, the original executes * Harmful prompts (hate, insults, sexual content, violence, and misconduct) are automatically blocked * No credits are charged for blocked prompts ![Prompt Results Panel](https://storage.crisp.chat/users/helpdesk/website/-/b/e/f/3/bef330df68156000/screenshot-2026-02-25-at-10284_1nezn8r.png) ## Results Panel ### Text Generation Results 1. Generated text appears immediately 2. Copy results with one click 3. “Start Over” resets the tool ![Text Generation Results](https://storage.crisp.chat/users/helpdesk/website/-/b/e/f/3/bef330df68156000/screenshot-2026-02-25-at-10282_wtquhc.png) ### Image Generation Results 1. Generated image displayed inline 2. Expand for larger view 3. Download full-size image **Note:** Generated images are stored temporarily and available for download for one hour after generation. ![Image Generation Results](https://storage.crisp.chat/users/helpdesk/website/-/b/e/f/3/bef330df68156000/screenshot-2026-02-25-at-10284_1ukpish.png) # Pricing & Credit Breakdown * Prompt Generation: 1 credit per prompt * Prompt Fine Tuning: 1 credit per tune * Prompt Execution: Pricing depends on the selected model, not prompt length. See the chart below for pricing: | 1 Credit Per Generated Text Response | 5 Credits Per Generated Text Response | 10 Credits Per Generated Text Response | 40 Credits Per Generated Image | | ---- | | [GPT-5 Nano](https://developers.openai.com/api/docs/models/gpt-5-nano), [Gemini 2.5 Flash Lite](https://deepmind.google/models/gemini/flash-lite/), [Amazon Nova Lite](https://aws.amazon.com/nova/?trk=33dc490e-0fb2-4cb1-a521-3941c13b64c0&sc_channel=ps&ef_id=CjwKCAiA2PrMBhA4EiwAwpHyC-qzY8ahdMckWVzRrnxg6AJXhld_NCFbigAJ_KZl3CIEX-5id-BjAxoCxTMQAvD_BwE:G:s&s_kwcid=AL!4422!3!795877020851!p!!g!!nova%20lite!23532472972!194311072244&gclid=CjwKCAiA2PrMBhA4EiwAwpHyC-qzY8ahdMckWVzRrnxg6AJXhld_NCFbigAJ_KZl3CIEX-5id-BjAxoCxTMQAvD_BwE), [AI21 Jamba 1.5 Mini,](https://www.ai21.com/jamba/) [Mistral Small](https://mistral.ai/news/mistral-small-3-1), [Anthropic Claude 3 Haiku ](https://www.anthropic.com/news/claude-3-haiku) | [GPT-5 Mini](https://developers.openai.com/api/docs/models/gpt-5-mini), [Gemini 2.5 Flash](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-flash), [Amazon Nova Pro](https://aws.amazon.com/nova/?trk=33dc490e-0fb2-4cb1-a521-3941c13b64c0&sc_channel=ps&ef_id=CjwKCAiA2PrMBhA4EiwAwpHyC6_SCTBBM63EOl5_eTZ360eDla1cErvmUUP_zushA4_ignT93WEyXBoCgegQAvD_BwE:G:s&s_kwcid=AL!4422!3!795877020851!p!!g!!amazon%20nova!23532472972!194311072244&gclid=CjwKCAiA2PrMBhA4EiwAwpHyC6_SCTBBM63EOl5_eTZ360eDla1cErvmUUP_zushA4_ignT93WEyXBoCgegQAvD_BwE), [AI21 Jamba 1.5 Large](https://www.ai21.com/jamba/), [Mistral Large](https://mistral.ai/news/mistral-large) | [Cohere Command R+](https://docs.cohere.com/docs/command-r-plus), [Claude 3 Sonnet](https://claude.ai/login), [Claude 3.5 Sonnet ](https://www.anthropic.com/news/claude-3-5-sonnet) | [Amazon Nova Canvas](https://nova.amazon.com/imagegeneration), [Amazon Titan Image Generator V2](https://aws.amazon.com/blogs/aws/amazon-titan-image-generator-v2-is-now-available-in-amazon-bedrock/?refid=33dc490e-0fb2-4cb1-a521-3941c13b64c0), [Gemini 2.5 Flash Image ](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-flash-image) | # Frequently Asked Questions ### Why did generation or tuning fail? Prompt generation and tuning rely on external LLM providers. Temporary outages may impact performance. ### Why did my prompt fail to execute? Possible reasons: * The selected model is temporarily unavailable * The prompt was flagged by harmful content filters Blocked prompts are **not charged**. ### Can I execute on non-recommended models? No. Execution is limited to the two recommended models displayed. ### Are my prompts stored? No. If you close or refresh the page: * Generated prompts are lost * Results are not saved * Images expire after one hour Be sure to copy or download your outputs.