We need to output JSON with fields: joke_text, headline, meme_caption, image_pro...

We need to output JSON with fields: joke_text, headline, meme_caption, image_prompt, meta_title, meta_description. Must be in German language for joke_text, headline, meme_caption, but image_prompt must be English. Must start with "Chuck Norris" or mention him. Must be unique, not repeating existing jokes. Must be funny, classic Chuck Norris style. Must avoid profanity, politics, real suffering. Must be creative. Provide JSON only, no markdown code blocks. We need to ensure each field length constraints: headline max 80 characters, meme_caption max 80 characters, meta_title max 60 characters, meta_description max 160 characters. We need to craft a joke that starts with "Chuck Norris". Something like: "Chuck Norris kann jeden Schneemann zum Leben erwecken, indem er einfach einen Blick darauf wirft." But need to be unique. Maybe: "Chuck Norris kann einen Schneemann im Sommer zum Schmelzen bringen, ohne dass es warm wird." That's okay. But need to be classic Chuck Norris style: exaggeration, impossible feats. Let's craft: "Chuck Norris kann einen Kühlschrank öffnen, ohne dass das Licht ausgeht." That's a known variant but maybe not unique. Let's think of something fresh. Maybe: "Chuck Norris kann einen Schneemann im Winter zum Schießen bringen, und der Schuss ist immer ein Treffer, weil er nie einen Zielverfehlungsfehler macht." Hmm. Better: "Chuck Norris kann einen Kaugummi aus einem Gummibaum pflücken, ohne dass der Baum protestiert." Not great. We need something that fits the image prompt: a hyper-realistic portrait of a tough legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build, black sleeveless vest. The scene from the joke must be described in the image_prompt. So we need to embed a specific scene from the joke into the image_prompt. For example, if the joke is about Chuck Norris lifting a car with one hand while drinking coffee, the image prompt should describe that scene. Thus we need to decide a joke that can be visualized: maybe Chuck Norris is standing on a mountain of broken keyboards, typing with one finger, causing the computer to crash. Or Chuck Norris is playing a piano with his beard. But we need to keep it classic: maybe "Chuck Norris kann einen Kühlschrank mit einem Blick zum Schweigen bringen, sodass die Milch nie mehr abläuft." Not visual. Better: "Chuck Norris kann einen Schneemann im Sommer zum Schmelzen bringen, ohne dass es warm wird." That's not visual. Maybe we can do: "Chuck Norris kann einen Kaugummi aus einem Baum pflücken, und der Baum wirft ihm einen Applaus zu." Not visual. We need a scene that can be described: maybe "Chuck Norris hält einen Laptop in einer Hand, während er mit dem Daumen einen Feuerball erzeugt." That could be visualized. But we need to keep it simple and classic: maybe "Chuck Norris kann einen Schneemann im Winter zum Schießen bringen, und der Schneemann schreit 'Ich bin ein Schneemann!'" Not good. Let's think of a typical Chuck Norris joke: "Chuck Norris can divide by zero." That's English. In German: "Chuck Norris kann durch Null teilen, und das Ergebnis ist immer 42." That's a math joke. But we need to be unique and not repeat existing ones. The "divide by zero" is common. Let's avoid that. Maybe: "Chuck Norris kann einen Computer Virus mit einem einzigen Blick in Schach halten, sodass er sich selbst löscht." That could be visualized: Chuck Norris staring at a computer screen causing the virus to disappear. But we need a scene: "Chuck Norris stares at a computer screen, and the virus evaporates into pixels." That can be described. Let's craft a joke: "Chuck Norris kann einen Computer-Virus mit einem einzigen Blick zum Schweigen bringen, sodass er sich selbst in Luft auflöst." That is unique. Now we need to produce JSON fields: - joke_text: the full joke text in German, starting with "Chuck Norris". Must be a single string. - headline: short title max 80 chars. - meme_caption: image caption max 80 chars. - image_prompt: English prompt for image generation, describing the specific scene from the joke, with the required style. Must include the scene. - meta_title: SEO Title max 60 chars. - meta_description: SEO Description max 160 chars. We must ensure lengths. Let's craft: joke_text: "Chuck Norris kann einen Computer-Virus mit einem einzigen Blick zum Schweigen bringen, sodass er sich selbst in Luft auflöst." That's 108 characters? Let's count roughly: "Chuck Norris kann einen Computer-Virus mit einem einzigen Blick zum Schweigen bringen, sodass er sich selbst in Luft auflöst." That's okay. Now headline: maybe "Chuck Norris vs. Computer-Virus" (max 80). That's 20 chars. meme_caption: maybe "Wenn Chuck Norris einen Virus sieht, wird er zum Staub." That's 44 chars. image_prompt: Must be English, describing the scene: "A hyper-realistic portrait of a tough, legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build, black sleeveless vest, standing in a dimly lit server room, holding a glowing computer monitor that is cracking as a virus dissolves into pixels, dramatic cinematic lighting, American flag background, photorealistic action movie poster, no text, no captions." We need to include the specific scene: "holding a glowing computer monitor that is cracking as a virus dissolves into pixels". That fits. Now meta_title: SEO Title max 60 chars. Something like "Chuck Norris besiegt Computer-Viren mit einem Blick". Count characters: "Chuck Norris besiegt Computer-Viren mit einem Blick" length? Let's count: Chuck(5) + space(1)=6, Norris(6)=12, space=13, besiegt(7)=20, space=21, Computer-Viren(14 inc hyphen?) Actually "Computer-Viren" is 13 inc hyphen? Let's count: C(1) o2 m3 p4 u5 t6 e7 r8 -9 V10 i11 r12 e13 n14. So 14 characters. So far 21+14=35, space=36, mit(3)=39, space=40, einem(5)=45, space=46, Blick(5)=51. So total 51 characters, within 60. Good. meta_description: SEO Description max 160 chars. Something