LTX 2.5 对白与人声提示词
收录 80 多条 LTX-2.5 视频提示词,来自 Civitai、fal.ai 的 LTX-2.5 示例页和 X。每条都逐字全文收录、署名作者、链回原帖,原帖有视频的,卡片里直接播放。这些提示词都不是我们写的。Civitai 上的大多是作者在自己电脑上用 LTX-2.5 开源权重跑的(ComfyUI 里的 dev 或蒸馏版),模型是同一个家族,但没有经过我们的 API。
最近更新: · LTX 2.5 Fast API 价格与在线体验 · API 文档
一句话摘要
LTX-2.5 是 Lightricks 在 2026 年 8 月 11 日发布的开源权重视频模型,每条视频都会连声音一起生成。apimodels.app 上有两档:LTX 2.5 Fast(ltx-2.5-fast)可以只给提示词,也可以再加一张首帧图或一段音频,输出 720p、1080p 或 1440p,24 到 50 fps,6 到 20 秒,$0.081/秒起;LTX 2.5 Pro(ltx-2.5-pro)是 Lightricks 定位给成片用的一档,只做文生和首帧图生,720p 或 1080p,6、8、10 秒三档,$0.108/秒起。提示词最长 5000 字符;没有 4K、尾帧、参考图和视频输入;失败不扣费。本页收录 80 多条别人公开发布的 LTX-2.5 提示词,来自 Civitai、fal.ai 的 LTX-2.5 页面和 X。Civitai 上很多是在本地用开源权重、配合作者自己的参数跑的,同一段文字放到 API 上,出来的效果可能不一样。
出处:Lightricks LTX(模型方官方站点)
科幻咖啡馆里的吐槽
Use the provided start image exactly as the first frame. Short clip of a woman in a sci-fi cafe with holographic waitresses and futuristic flying cars outside. Performance: The woman speaks and gesticulates casually, saying "No, it's not that. I love my holo-phone. It's just that... well, okay, this is just an example, but you'd think by now we'd be done wiping our asses with paper!" Keep the acting natural and cinematic.
阴天户外采访
A cinematic outdoor interview scene in soft overcast daylight. The camera begins in a medium over-the-shoulder shot from behind the male interviewer, keeping the blonde woman in sharp focus while he holds the microphone naturally toward her. She maintains eye contact with him and speaks calmly, her lips moving naturally in sync with her words: "I think people sometimes forget how much the little things matter." She gives a small thoughtful smile and makes a subtle hand gesture as she finishes. The camera slowly pushes closer toward her face with shallow depth of field. She pauses, listening attentively. The interviewer slightly lowers the microphone toward himself and asks in a relaxed, curious voice, "So what made you realize that?" The camera smoothly shifts focus toward him for a brief reaction shot, then returns to the woman as the microphone moves naturally back toward her. She takes a short breath, tilts her head slightly and replies with a warmer expression, "Honestly? Someone reminded me when I needed it most." She gives a gentle smile and briefly looks down before meeting his eyes again. The camera settles into a close-up of her face for the final moment, capturing subtle blinking, breathing, loose strands of hair moving in the breeze and natural facial micro-expressions. Realistic conversational timing, natural pauses between speakers, accurate lip synchronization, restrained body language, smooth cinematic camera movement, natural motion blur, realistic skin texture, shallow depth of field, soft ambient city noise, distant birds, light wind through nearby trees, no music.
商店退货婚纱
She speaks to the assistant off-camera, polite and matter-of-fact. "I'd like to return this." She glances at the bag. "I bought it for my sister's wedding, which... is now my wedding... so... I'll need it in white." She smiles bashfully. Quiet store, a soft till beep. Single continuous take.
工地女子受访谈家族
She speaks loudly over nearby construction noise to the interviewer off-camera, composed and pleased. "I know it's construction, but I run this site like a family." She pauses. "Which means nobody talks, and one of us isn't allowed at Christmas." A small satisfied nod. She holds it. Sound of power tools, diesel engines, a reversing beeper. Single continuous take.
侦探分析冷静嫌疑人
The detective speaks to her colleague off-camera, grave and authoritative. "We're looking for someone calm. Methodical." She pauses, looks down at the notebook. "They rinsed the mug. Then they left it in the sink. So, calm, but not raised right." She looks back up and holds it. Quiet room, a faint police radio. Single continuous take.
新西兰女子求职面试
She speaks with a native Kiwi accent to the interviewer off-camera, pleasant and entirely sincere. "Why here?" A small nod, as if it's a good question. "I'd be keen to work wherever, honestly. This is the closest Woolies to my house." She pauses. "There's a Bunnings closer. But they haven't called back." She holds the look, helpful. Quiet office, a distant printer. Single continuous take.
渔女谈论水中鱼
She speaks to the interviewer off-camera, slow and quietly certain. "Grew up on boats. And I've learned one thing." She pauses, looks out at the water. "The fish are in the water. Not on the boat. You'd be surprised how many people check the boat." She looks back. Gulls, water slapping the hull. Single continuous take.
机长广播飞行趣事
She speaks into the handset, warm and practised, like every announcement she's ever made. "Good morning, this is your captain. Flight time today is about an hour." She pauses. "We'll be cruising at... hmm... four hundred feet, I guess." A beat. "I like to see things." She sets the handset down. Cabin hum, a seatbelt chime. Single continuous take.
两个哥布林守卫闲聊
The goblin on the right says, "Aren't you cute?" The goblin on the left says, "Yup, we got jobs here as guards."
咖啡馆经理冷面受访
She speaks to the interviewer off-camera, calm and quietly certain. "How long have I worked here? Since Tuesday." She pauses. "I'm the manager now. Sarah was the manager. Sarah went to lunch." A small nod. She holds it. Café room tone, the hiss of a steam wand. Single continuous take.
阿姆斯特丹夜色自拍
A spunky 20-year-old African woman with curly brown hair, wearing oversized tortoiseshell sunglasses and a denim jacket, walks through the vibrant, neon-lit streets of Amsterdam’s nightlife district at night. The camera follows her in a handheld selfie style, keeping her centered as she strolls past bustling crowds, glowing street lamps, and glowing shop signs reflecting off wet pavement. She beams with wide, joyful smiles, laughing as she speaks to the camera. Her voice is bright and playful. Shot 1: She begins walking forward, smiling and talking directly to the camera, exclaiming “This is my first time in Amsterdam and I already love it!” The camera gently bobs with her steps as she laughs, revealing the lively street behind her with trams passing and people mingling. She gestures with her free hand, excitement radiating in her posture. Shot 2: As she continues walking, she pauses slightly, turning her head to glance around playfully, then leans in toward the camera with a mischievous grin and says, “I already tried some weed in a local coffee shop, and I met a cute guy there.” She raises her eyebrows and winks, the camera tilting slightly to catch the sparkle in her eyes. Shot 3: She leans even closer, grinning, then whispers teasingly, “I’ll totally go home with that guy, and just see what will happen.” Her cheeks flush slightly as she says this, her eyes darting away for a split second before returning with an even bigger grin. She chuckles nervously, “Oh my gosh, I’m so embarrassed saying that,” and shakes her head with a giggle, her hand fluttering near her sunglasses. Behind her, the street noise fades slightly as she pauses, soaking in the moment, the neon lights casting colorful glows across her face. The video ends with her stepping forward again, laughing heartily, the camera following her as she walks away, still smiling, the energy of the city pulsing behind her. No fade to black—just her motion and laughter carrying the vibrant rhythm of Amsterdam’s nightlife.
爱尔兰女子对镜说话
Create a SFW video where the woman, 25yrs, Irish, speaking English with lilting Irish accent, says: "Hi guys, Coral here. I understand someone on this site is asking for my number. I'm flattered, but I'm a shy gal, so, sorry. But hey, from my portfolio you can see I frequent the MTA, the Underground and the DART & LUAS in Dublin. If you're on one of them, stay alert and maybe, who knows?"
1974年维也纳旅馆对峙
A slow dolly in on a man seated at a small table in a Vienna hotel room in 1974, framed from just off his left shoulder. He is in his fifties, thinning grey hair, tie loosened, a cigarette burning down in an ashtray he has forgotten about. A single practical lamp on the table lights him hard from below the eyeline and leaves the rest of the room in near darkness, the window behind him carrying wet streetlight and nothing else. He looks up at someone off camera, waits, and says quietly in English with a slight German accent: "You had the file for six weeks and you brought me a photograph." Then he taps the ash and looks back down. 40mm at f/2, cool tungsten falloff, fine grain, one continuous shot with no cuts. Sound: his voice close and dry with the room's small reverb under it, traffic hissing on wet road below, a radiator ticking, no music.
女律师的结案陈词
A woman in her forties speaking American English, low and measured, a defense lawyer near the end of a long trial, no urgency anywhere in the delivery, quiet courtroom room tone sitting behind the voice: "You have heard eleven days of testimony. Not one of those days put my client in that room." 9 seconds.
怎么写好 LTX 2.5 提示词
- 1一个镜头写成一段话,按发生顺序写. 本页 fal.ai 的那些示例,每条都是从头到尾描述一个机位:景别和主体、发生了什么、镜头怎么动、光线,最后才是镜头焦段。大意是「缓慢推近擂台上的拳击手,顶光穿过烟雾打在肩上,50mm、f/2」。一串关键词堆在一起,模型拿不到先后顺序。
- 2把声音写进去,反正每条都会出声. LTX-2.5 的声音和画面一起生成,我们的 API 也没有静音选项,所以要不要声音不由你,怎么响可以由你。台词放进引号、写明是谁说的,再写环境音和最关键的那一下声音。本页有几条 Civitai 提示词把两个人的对话一句一句写好,每个人开口前先交代嗓音。
- 3动作长度要对得上 API 的时长档. Fast 的文生和图生是 6 到 20 秒、只能取偶数秒,1440p 或 48 / 50 fps 时只有 6、8、10 秒;Pro 只有 6、8、10 秒。本地工作流常用帧数定长度,除以帧率就知道原帖要的是几秒,再挑最接近的一档。API 不会看提示词自己定长度,要自己传 duration。
- 4传了首帧,就写动作,别再描述画面. 图生视频里,主体、场景和构图已经由那张图定死了,提示词就专心写什么在动:谁转身、镜头怎么走、听到什么。成片会保持图片的比例;带首帧又传 aspect_ratio,会直接返回 400。
- 5音频生视频:音轨说了算,提示词负责选角. LTX 2.5 Fast 收到 audio_url 后,会把这段音频原样当成片的声音,再按它生成画面。提示词写清楚画面里是谁、怎么动、怎么取景就够了,比如我们文档里的例子:明亮的演播室里,一位年轻主持人对着镜头讲话,手势自然,中景。长度默认跟着音频走,也可以自己设 duration。Pro 不收音频。
- 6API 收不了的东西,先删掉. 本地 LTX-2.5 工作流里可以挂负面提示词、LoRA 触发词、额外的引导图或尾帧。我们的请求只收提示词、一张首帧、一段音频和生成参数,别的都不收:尾帧、参考图、视频输入会返回 400。也没有负面提示词这个字段,原帖写在负面提示词里的东西,不挪进正文就等于没了。
- 7画面文字和很乱的物理,别指望一次成. Lightricks 自己的提示词指南也说了,画面里的文字和混乱的物理运动目前还不稳定。如果一个镜头靠一块看得清的招牌,或者一堆东西互相撞,做好多跑几次的准备,文字也可以剪辑时再加。
提示词公式
景别和主体 + 按先后写的一个清楚动作 + 镜头运动 + 光线和时间 + 焦段或胶片质感 + 引号里的台词和环境音 + 这一档允许的时长(用 duration 传,不写进提示词)
例(本库收录的原文,原样引用,作者 hichuicomfy,发在 Civitai):Style: high-resolution food photography - close-up - The chocolate is poured over the whipped cream and berries. The chocolate flows slowly, creating rich, glossy streaks across the white cream and the blue blueberry, the deep brown contrasting with the vibrant fruit. A soft, rich sound of chocolate dripping and settling accompanies the pour.
常见问题
LTX-2.5 是什么?
LTX-2.5 是 Lightricks 的视频模型,2026 年 8 月 11 日以开源权重发布,每条视频都会把声音和画面一起生成。Lightricks 的 API 分两档:Fast 主打速度和成本,Pro 按它的说法主打画面保真和运动稳定。Lightricks 还说 LTX-2.5 新增了原生多镜头、新的视频解码器和更好的提示词遵循,这些是厂商的说法,不是我们的测试结论。
这些提示词是谁写的?
都是别人写的:每一条都是创作者在 Civitai 或 X 上公开发的,或者是 fal.ai 放在它 LTX-2.5 页面上的,我们一个字都没写、也没改。每张卡片都逐字全文引用提示词、写明作者、链回原帖。提示词和视频归原作者所有;我们自己写的说明和整理的数据按 CC BY 4.0 授权。要更正署名或撤下某一条,发邮件给我们,或者在 GitHub 仓库提 issue。
Civitai 上的提示词,我用 API 跑出来不一样,为什么?
因为本页 Civitai 的提示词大多是作者在自己电脑上用 LTX-2.5 开源权重跑的,不是在我们的 API 上,同一段文字出来的效果自然会有差别。本地 ComfyUI 可能用的是 dev 或蒸馏版权重,还可能挂了 LoRA、调了采样参数、写了负面提示词或者加了引导图,这些都不在提示词文字里。文字能照搬,这些设置搬不过来。卡片上的视频是作者那套设置的结果,在 Fast 或 Pro 上要多试几次才能接近。
在哪能跑这些提示词?多少钱?
在 apimodels.app 上,模型名 ltx-2.5-fast 和 ltx-2.5-pro,接口 POST /v1/video/generations,按成片秒数计费。Fast:720p $0.081/秒、1080p $0.117/秒、1440p $0.171/秒,720p 6 秒 $0.486,1080p 10 秒 $1.17,1080p 20 秒 $2.34。Pro:720p $0.108/秒、1080p $0.153/秒,1080p 10 秒 $1.53。两档都比 Lightricks 官方 API 价低 10%;Fast 的音频生视频按成片长度计到 0.01 秒;失败不扣费。常见的用法是先在 Fast 上试提示词,选中的镜头再拿到 Pro 上出,两档字段一样,只改一个模型名。
本页的提示词放到 API 上被拒或者失败,为什么?
超出 API 限制的请求,会在建任务之前直接返回 400,不扣费。限制有这些:提示词最长 5000 字符;Fast 是 6 到 20 秒的偶数秒,Pro 只有 6、8、10 秒;没有 4K、尾帧、参考图和视频输入;Pro 不收音频、没有 1440p;传了首帧就不能再传 aspect_ratio。任务开始以后,也可能因为内容审核或上游繁忙而失败,同样不扣费。
LTX-2.5 会自己配声音吗?
会:每条视频的声音都和画面一起生成,我们的 API 也没有静音选项。在意声音的话,把台词放进引号,写明环境音。用 LTX 2.5 Fast 还可以直接传你自己的音轨(audio_url),画面会按这段音频来生成。
这些提示词和视频能商用吗?
提示词和视频归原作者所有,商用前先看原帖怎么说。你自己生成的视频,按你用来生成的那个服务的条款执行。如果你在本地跑开源权重,它用的是 Lightricks 的 LTX-2.x Community License,有自己的条件,商用前先读一遍。
在 apimodels.app 上跑这些提示词
LTX 2.5 Fast 已经上线,和另外 180+ 个图像、视频、语音、大模型共用一个 API、一个 key,按量计费、注册送 $0.10 试用额度。
这份提示词库在 GitHub 上是开源的
整份合集以 Markdown 和 JSON 放在 GitHub 上:每条提示词都带作者署名和原帖链接(Civitai、fal.ai 或 X)。提示词逐字全文引用,归原作者所有。视频也归作者所有,只链原帖、不转存。我们自己写的说明和整理的数据按 CC BY 4.0 授权。要更正署名或撤下某一条,提 issue 即可。
在 GitHub 上查看