A social post prompt builder removes the guesswork from getting an LLM to write captions that actually fit each platform. A LinkedIn post and a TikTok caption are not interchangeable — the hook, length, emoji density, and hashtag strategy all differ — yet most people paste the same vague request into a chatbot and get bland, channel-agnostic text. This tool encodes the conventions of each network into a single optimised prompt you can reuse for every post.
How it works
You select the platform, which sets the hard constraints: character limits, line-break style, emoji and hashtag norms, and the kind of opening hook that performs there. You then describe the content type (announcement, tip, story, promotion), your brand voice in plain adjectives, and the call to action you want readers to take. The builder assembles these into a structured instruction block with a placeholder for your topic, plus rules that tell the model to respect the platform’s length and to avoid generic marketing clichés. Everything is generated locally in your browser — no topic or brand data is uploaded.
Tips and examples
Be specific in the brand voice field: “warm, plain-spoken, lightly witty, never salesy” produces far better output than “professional”. For X, keep your call to action implicit or very short, since the character budget is tight. For Instagram and TikTok, the hook in the first line is what stops the scroll, so the prompt asks the model to lead with it. Save your favourite generated prompt as a snippet and only swap the topic each time — this keeps a brand’s voice consistent across dozens of posts and across team members who all run the same template.