Turn rough ideas into structured prompts for ChatGPT, Claude, Gemini, image and video AI. Free.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Turns a rough idea ("cold email to a cafe owner", "a cat astronaut poster") into a detailed, structured prompt ready to paste into an AI tool. Returns the prompt text.
What the prompt should get done, in the user's words. Include audience, tone, length or format if the user gave them.
text for chat assistants (ChatGPT, Claude, Gemini), image for image models (Midjourney, Nano Banana, Flux, GPT Image), video for video models (Sora, Veo, Kling, Runway).
Text prompts only. standard picks the structure from written_for. reasoning is for step-by-step problem solving. The rest are named prompt frameworks (RACE, CARE, APE, CREATE, TAG, CREO, RISE, PAIN, COAST, ROSES, RESEE).
Text prompts only. chatgpt follows OpenAI prompting guidance, claude follows Anthropic guidance, any uses Feedough's ORACLE structure (Outcome, Role, Audience, Constraints, Layout, Evidence) that works everywhere.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http prompt-generator https://www.feedough.com/mcp
Turns a rough idea ("cold email to a cafe owner", "a cat astronaut poster") into a detailed, structured prompt ready to paste into an AI tool. Returns the prompt text.
| Argument | Type | Description |
|---|---|---|
| idea* | string | What the prompt should get done, in the user's words. Include audience, tone, length or format if the user gave them. |
| kind | string | text for chat assistants (ChatGPT, Claude, Gemini), image for image models (Midjourney, Nano Banana, Flux, GPT Image), video for video models (Sora, Veo, Kling, Runway). |
| framework | string | Text prompts only. standard picks the structure from written_for. reasoning is for step-by-step problem solving. The rest are named prompt frameworks (RACE, CARE, APE, CREATE, TAG, CREO, RISE, PAIN, COAST, ROSES, RESEE). |
| written_for | string | Text prompts only. chatgpt follows OpenAI prompting guidance, claude follows Anthropic guidance, any uses Feedough's ORACLE structure (Outcome, Role, Audience, Constraints, Layout, Evidence) that works everywhere. |
{
"type": "object",
"required": [
"idea"
],
"properties": {
"idea": {
"type": "string",
"description": "What the prompt should get done, in the user's words. Include audience, tone, length or format if the user gave them."
},
"kind": {
"enum": [
"text",
"image",
"video"
],
"type": "string",
"default": "text",
"description": "text for chat assistants (ChatGPT, Claude, Gemini), image for image models (Midjourney, Nano Banana, Flux, GPT Image), video for video models (Sora, Veo, Kling, Runway)."
},
"framework": {
"enum": [
"standard",
"reasoning",
"race",
"care",
"ape",
"create",
"tag",
"creo",
"rise",
"pain",
"coast",
"roses",
"resee"
],
"type": "string",
"default": "standard",
"description": "Text prompts only. standard picks the structure from written_for. reasoning is for step-by-step problem solving. The rest are named prompt frameworks (RACE, CARE, APE, CREATE, TAG, CREO, RISE, PAIN, COAST, ROSES, RESEE)."
},
"written_for": {
"enum": [
"any",
"chatgpt",
"claude"
],
"type": "string",
"default": "any",
"description": "Text prompts only. chatgpt follows OpenAI prompting guidance, claude follows Anthropic guidance, any uses Feedough's ORACLE structure (Outcome, Role, Audience, Constraints, Layout, Evidence) that works everywhere."
}
},
"additionalProperties": false
}Rewrites an existing prompt with the change the user asks for ("shorter", "add three examples", "for a B2B audience"). Returns the new prompt.
| Argument | Type | Description |
|---|---|---|
| idea | string | Optional: the original idea the prompt was written for. |
| kind | string | text for chat assistants (ChatGPT, Claude, Gemini), image for image models (Midjourney, Nano Banana, Flux, GPT Image), video for video models (Sora, Veo, Kling, Runway). |
| change* | string | What should be different, in the user's words. |
| prompt* | string | The prompt to improve, in full. |
| framework | string | Text prompts only. standard picks the structure from written_for. reasoning is for step-by-step problem solving. The rest are named prompt frameworks (RACE, CARE, APE, CREATE, TAG, CREO, RISE, PAIN, COAST, ROSES, RESEE). |
| written_for | string | Text prompts only. chatgpt follows OpenAI prompting guidance, claude follows Anthropic guidance, any uses Feedough's ORACLE structure (Outcome, Role, Audience, Constraints, Layout, Evidence) that works everywhere. |
{
"type": "object",
"required": [
"prompt",
"change"
],
"properties": {
"idea": {
"type": "string",
"description": "Optional: the original idea the prompt was written for."
},
"kind": {
"enum": [
"text",
"image",
"video"
],
"type": "string",
"default": "text",
"description": "text for chat assistants (ChatGPT, Claude, Gemini), image for image models (Midjourney, Nano Banana, Flux, GPT Image), video for video models (Sora, Veo, Kling, Runway)."
},
"change": {
"type": "string",
"description": "What should be different, in the user's words."
},
"prompt": {
"type": "string",
"description": "The prompt to improve, in full."
},
"framework": {
"enum": [
"standard",
"reasoning",
"race",
"care",
"ape",
"create",
"tag",
"creo",
"rise",
"pain",
"coast",
"roses",
"resee"
],
"type": "string",
"default": "standard",
"description": "Text prompts only. standard picks the structure from written_for. reasoning is for step-by-step problem solving. The rest are named prompt frameworks (RACE, CARE, APE, CREATE, TAG, CREO, RISE, PAIN, COAST, ROSES, RESEE)."
},
"written_for": {
"enum": [
"any",
"chatgpt",
"claude"
],
"type": "string",
"default": "any",
"description": "Text prompts only. chatgpt follows OpenAI prompting guidance, claude follows Anthropic guidance, any uses Feedough's ORACLE structure (Outcome, Role, Audience, Constraints, Layout, Evidence) that works everywhere."
}
},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 5, 2026.