generate_image

Instructions

Implementation Reference

• src/tools.ts:217-231 (registration)

  The 'generate_image' tool is registered using server.tool() in the registerTools function. It defines the tool name, description, input schema (prompt, aspect_ratio, model), and the handler callback.

  server.tool(
    'generate_image',
    'Generate an AI image from a text prompt on PostEverywhere. The image is saved to your media library and can be attached to posts via media_ids. Choose from 4 models: nano-banana-pro (best quality, 15 credits), ideogram-v2 (best for text-in-image, 8 credits), gemini-3-pro (balanced, 5 credits), flux-schnell (fastest, 1 credit). Requires the "ai" scope on your API key.',
    {
      prompt: z.string().max(2000).describe('Text description of the image to generate'),
      aspect_ratio: z.enum(['1:1', '16:9', '9:16', '4:3', '3:4', '4:5', '5:4']).optional().default('1:1').describe('Aspect ratio for the generated image'),
      model: z.enum(['nano-banana-pro', 'ideogram-v2', 'gemini-3-pro', 'flux-schnell']).optional().default('nano-banana-pro').describe('AI model to use for generation'),
    },
    async ({ prompt, aspect_ratio, model }) => {
      const result = await client.generateImage({ prompt, aspect_ratio, model });
      return {
        content: [{ type: 'text' as const, text: JSON.stringify(result, null, 2) }],
      };
    }
  );

• src/tools.ts:225-230 (handler)

  The handler function for 'generate_image' that receives {prompt, aspect_ratio, model} and delegates to client.generateImage(), returning the JSON-stringified result.

  async ({ prompt, aspect_ratio, model }) => {
    const result = await client.generateImage({ prompt, aspect_ratio, model });
    return {
      content: [{ type: 'text' as const, text: JSON.stringify(result, null, 2) }],
    };
  }

• src/tools.ts:220-224 (schema)

  Input schema defined with Zod: prompt (string, max 2000), aspect_ratio (enum, optional, default '1:1'), model (enum, optional, default 'nano-banana-pro').

  {
    prompt: z.string().max(2000).describe('Text description of the image to generate'),
    aspect_ratio: z.enum(['1:1', '16:9', '9:16', '4:3', '3:4', '4:5', '5:4']).optional().default('1:1').describe('Aspect ratio for the generated image'),
    model: z.enum(['nano-banana-pro', 'ideogram-v2', 'gemini-3-pro', 'flux-schnell']).optional().default('nano-banana-pro').describe('AI model to use for generation'),
  },

• src/client.ts:212-218 (helper)

  The client method generateImage() on PostEverywhereClient that sends a POST request to /ai/generate-image with the prompt, aspect_ratio, and model body. Returns a Promise with media_id, model, aspect_ratio, credits_used, credits_remaining, and message.

  async generateImage(body: {
    prompt: string;
    aspect_ratio?: string;
    model?: string;
  }): Promise<{ media_id: string; model: string; aspect_ratio: string; credits_used: number; credits_remaining: number; message: string }> {
    return this.request('POST', '/ai/generate-image', body);
  }