{"repo": "bu-bu-xxx/video-chain", "files": {"GOAL.md": {"5afa85a1f88de9a6116bdcb02807a77a199751dd": {"t": 1785768965.1595068, "n": 2281, "h": "acf08b0fe651964d", "c": "我想实现一个workflow：\n用户输入图片和文字描述需求，根据用户需求生成每个视频片段的文字prompt，通过sora2 api输入文字prompt，并且输入上一个视频的最后帧画面作为图片prompt输入，进行每个视频片段的生成，以确保视频的coherent，最后拼接成一个长视频输出\n\ndevelopment requriement:\n1. 我们会使用`aihubmix`提供的sora2模型和gemini-3-pro-preview模型的api实现workflow\n   \n2. 所有提示词用中文\n   \n3. `aihubmix` `sora2` 模型调用方式Example：\n\nGenerate With a Reference Image:\n```python\nimport requests\n\nurl = \"https://aihubmix.com/v1/videos\"\n\nheaders = {\n    \"Authorization\": \"Bearer sk-***\"\n}\n\n\n# Prepare the multipart/form-data payload\nfiles = {\n    \"prompt\": (None, \"The kitten is taking a nap under the tree.\"),\n    \"model\": (None, \"sora-2-pro\"),\n    \"size\": (None, \"1280x720\"),\n    \"seconds\": (None, \"4\"),\n\n    \"input_reference\": (\n        \"cat.jpeg\",\n        open(\"cat.jpeg\", \"rb\"),\n        \"image/jpeg\"\n    )\n}\n\nresponse = requests.post(url, headers=headers, files=files)\n\nprint(\"Status:\", response.status_code)\nprint(\"Response:\", response.text)\n```\n\nGenerate Without a Reference Image:\n```python\nimport requests\nimport json\n\nurl = \"https://aihubmix.com/v1/videos\"\nheaders = {\n    \"Content-Type\": \"application/json\",\n    \"Authorization\": \"Bearer sk-***\"\n}\n\npayload = {\n    \"model\": \"sora-2\",\n    \"prompt\": \"A astronaut floating in space, holding a glowing plant, with Earth and stars in the background, slow-motion movement\",\n    \"size\": \"720x1280\",\n    \"seconds\": '4',\n}\n\nresponse = requests.post(url, headers=headers, json=payload)\n\nprint(\"Status Code:\", response.status_code)\nprint(\"Response:\", response.text)\n```\n\n\n4. `aihubmix` `gemini-3-pro-preview` 模型调用方式Example：\n```python\nfrom openai import OpenAI\n\nclient = OpenAI(\n    api_key=\"AIHUBMIX_API_KEY\", # Your Key \"sk-***\"\n    base_url=\"https://aihubmix.com/v1\"\n)\n\nresponse = client.responses.create(\n    model=\"gemini-3-pro-preview\", \n    input=[\n        {\n            \"role\": \"user\",\n            \"content\": [\n                { \"type\": \"input_text\", \"text\": \"what is in this image?\" },\n                {\n                    \"type\": \"input_image\",\n                    \"image_url\": \"https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg\"\n                }\n            ]\n        }\n    ]\n)\n\nprint(response)\n```\n\n4. 先开发一个command line方式运行这个workflow，然后开发一个web界面可以运行以及展示\n"}}}}