You never have to reshoot a video again. Swap Character puts someone new in your clip, face and outfit included, while the place, motion, camera and timing stay exactly as they were. Add a photo of who you want, and the clip is remade around them. Find it under Instant Video on BLOCKLORE.

Change who is in the video. Keep everything else. Swap character in Video remix replaces the people in your clip with anyone you choose. The motion, camera and timing stay exactly the same. Upload a video, add a photo for each person you want to replace, and the rest of the shot carries over. Find it under Instant Video on BLOCKLORE.

AI music is getting a lot more usable, and now voice comes with it. Music: Suno v6 is the flagship, reliable and precise, recommended for most songs. Suno v6 Wild is experimental, more varied and unpredictable. Suno v6 Mini is faster. Voice: Qwen Audio 3.1 TTS Flash is real-time with 68 voices, voice cloning and voice design, multiple languages and dialects, and control over emotion, tone, speed and volume. Kling Text to Speech handles short text with a choice of voices, up to 120 characters per request. Kling Lip Sync matches a character’s mouth to uploaded audio or generated speech. Coming soon on BLOCKLORE.

Video generation now covers the whole process, from the first frame to the final edit. MiniMax H3 Max is a faster version of MiniMax H3. It does text-to-video and image-to-video, and you can set the first frame or the last frame. Veo 3.1, Fast, and Lite are joining the lineup. Vidu Q3 Ad Reference-to-Video makes commercial ads from 1 to 7 reference images and a prompt, in 720p or 1080p with synced audio. Vidu Q3 Drama Reference-to-Video takes a script plus your characters, scenes and props and builds a drama from them. Kling Video Effects gives you preset effects and stylized looks for social content. Kling Dual-Image Video Effects uses the same templates but blends two images into one animated clip. Got a long video? Smart Highlights · Cut pulls out the key moments. Smart Highlights · Narration adds an AI voiceover recap. Smart Highlights · Episode Summary recaps a whole episode. Coming soon on BLOCKLORE.

Making one good AI image is easy. Making ten that actually match is the hard part. Qwen-Image 2.1 Pro and Seedream 5.0 Flash both take up to 10 reference images and output transparent backgrounds. Seedream 5.0 Flash also separates layers and goes up to 2K, which suits ads and thumbnails. Qwen-Image 2.1 Pro is the one for product shots and marketing visuals. FLUX.1 Kontext Pro is the quick everyday editor. FLUX.1 Kontext Max follows edit instructions more precisely and renders text inside images more accurately, at the same speed. Nano Banana 2.1 is in the batch too. Coming soon on BLOCKLORE.

Before any video gets made, someone has to write the script. 19 language models are coming to BLOCKLORE. Claude: Claude Opus 5.5, Claude Sonnet 5.5, Claude Fable 5.1, Claude Haiku 4.5, Claude Opus 4.5, Claude Sonnet 4.5. GPT: GPT 6 Luna, GPT 6 Sol, GPT 6.1 Sol, GPT 6 Astra. Gemini: Gemini 3.8 Flash, Gemini 3.7 Flash, Gemini Omni Flash. Grok: Grok 4.7 for coding, agentic tasks and knowledge work, and Grok Build 0.1, a fast agentic coding model. GLM: GLM 5.3, a 1M-token flagship that's strong at coding, agent tasks and cybersecurity, and GLM 5.3 Flash, an open-weight MIT-licensed model that takes text, image and video input Qwen: Qwen 3.8 Flash, fast and low-cost with hybrid reasoning and image understanding. Qwen 3.7 Flash, the previous generation, stays on the list too. Coming soon on BLOCKLORE.

Rate this clip 1-10 👇 Made entirely on BLOCKLORE. Base image, keyframes, video, audio. Reply with your score, then post your own and we'll rate it back.

For base images, your best bet is GPT Image 2.5. It's the best model we have for a reason. Lock in character angles, lighting, and style on the reference image before moving to video. Spending two minutes on a frame saves you hours of throwing away bad generations.

When a single shot keeps failing, like a character falling or fighting, stop re-rolling the exact same prompt over and over. Try this instead: 1.) Take the scene before it and storyboard the missing steps in between. 2.) Keyframe those intermediate reference images to lock in style, then animate small movements separately. That gives you enough coverage to make the sequence work.

Seedance native audio can actually be solid if your prompting is right. For quick short-form videos, skip ElevenLabs completely. Put explicit tone and dialogue directions straight into your prompt to get usable native audio. Save dedicated dubbing for long-form narrative projects where voice control matters. Prompt used: "Medium cinematic shot of a sleek, well-groomed dog sitting on a lavish velvet chair next to an ornate side table, quietly finishing a small crystal dish of caviar. Natural lighting, shallow depth of field, high detail, 4k. Audio Prompt / Dialogue: [Voice: Clear, natural female voiceover] "The dog ate the caviar.""