# AI Film & Video Tools > A directory of AI tools and releases for filmmakers and video creators, each linked to its primary source. ## Categories - [Video generation](https://aifilmhq.com/category/video-generation/): 24 entries - [Image](https://aifilmhq.com/category/image/): 9 entries - [Voice and audio](https://aifilmhq.com/category/voice-and-audio/): 22 entries - [Editing](https://aifilmhq.com/category/editing/): 13 entries - [Upscaling](https://aifilmhq.com/category/upscaling/): 6 entries - [Lip sync](https://aifilmhq.com/category/lip-sync/): 6 entries - [Workflows](https://aifilmhq.com/category/workflows/): 23 entries - [Local](https://aifilmhq.com/category/local/): Tools and models you run on your own machine, e.g. via ComfyUI, Forge, Pinokio or the CLI. 25 entries ## Tools - [MiniMax H3](https://aifilmhq.com/tool/minimax-h3/): 4 entries - [ComfyUI](https://aifilmhq.com/tool/comfyui/): 3 entries - [WanGP](https://aifilmhq.com/tool/wangp/): 3 entries - [Flick](https://aifilmhq.com/tool/flick/): 2 entries - [Adobe Premiere Generative Media Tool](https://aifilmhq.com/tool/adobe-premiere-generative-media-tool/): 1 entry - [BACH 1.0 Pro](https://aifilmhq.com/tool/bach-1-0-pro/): 1 entry - [Beeble SwitchX 2.0](https://aifilmhq.com/tool/beeble-switchx-2-0/): 1 entry - [Boreal](https://aifilmhq.com/tool/boreal/): 1 entry - [Bulbul V4 Flash](https://aifilmhq.com/tool/bulbul-v4-flash/): 1 entry - [ChatGPT Images 2.5](https://aifilmhq.com/tool/chatgpt-images-2-5/): 1 entry - [Concat](https://aifilmhq.com/tool/concat/): 1 entry - [drama3-h3-talking](https://aifilmhq.com/tool/drama3-h3-talking/): 1 entry - [Eleven v4](https://aifilmhq.com/tool/eleven-v4/): 1 entry - [ElevenLabs Instant Soundtracks](https://aifilmhq.com/tool/elevenlabs-instant-soundtracks/): 1 entry - [FastH3-Omni](https://aifilmhq.com/tool/fasth3-omni/): 1 entry - [Fish Audio Drama 3](https://aifilmhq.com/tool/fish-audio-drama-3/): 1 entry - [FLUX 3 Image Klein](https://aifilmhq.com/tool/flux-3-image-klein/): 1 entry - [Gemini 3.8 Flash TTS](https://aifilmhq.com/tool/gemini-3-8-flash-tts/): 1 entry - [Gnani Timbre v3.0](https://aifilmhq.com/tool/gnani-timbre-v3-0/): 1 entry - [Google Vids (Gemini Omni 1.1)](https://aifilmhq.com/tool/google-vids-gemini-omni-1-1/): 1 entry - [H3 Max](https://aifilmhq.com/tool/h3-max/): 1 entry - [h3-metal](https://aifilmhq.com/tool/h3-metal/): 1 entry - [HeyGen Video](https://aifilmhq.com/tool/heygen-video/): 1 entry - [Kandinsky 6.0 Video](https://aifilmhq.com/tool/kandinsky-6-0-video/): 1 entry - [Kling 4.0](https://aifilmhq.com/tool/kling-4-0/): 1 entry - [Krea 3](https://aifilmhq.com/tool/krea-3/): 1 entry - [LipDub 2.0](https://aifilmhq.com/tool/lipdub-2-0/): 1 entry - [LTX-2.5 Layout to Render](https://aifilmhq.com/tool/ltx-2-5-layout-to-render/): 1 entry - [Mago](https://aifilmhq.com/tool/mago/): 1 entry - [MAI-Image-2.6](https://aifilmhq.com/tool/mai-image-2-6/): 1 entry - [MAI-Voice-2.1](https://aifilmhq.com/tool/mai-voice-2-1/): 1 entry - [mflux](https://aifilmhq.com/tool/mflux/): 1 entry - [Mirelo SFX 2.0](https://aifilmhq.com/tool/mirelo-sfx-2-0/): 1 entry - [MOSS-Image](https://aifilmhq.com/tool/moss-image/): 1 entry - [Nari Qwen3-TTS](https://aifilmhq.com/tool/nari-qwen3-tts/): 1 entry - [Nvidia AI for Media](https://aifilmhq.com/tool/nvidia-ai-for-media/): 1 entry - [NVIDIA VFX Nodes](https://aifilmhq.com/tool/nvidia-vfx-nodes/): 1 entry - [OpenShot 4.0](https://aifilmhq.com/tool/openshot-4-0/): 1 entry - [Pika](https://aifilmhq.com/tool/pika/): 1 entry - [Plenio Music Production System](https://aifilmhq.com/tool/plenio-music-production-system/): 1 entry - [Runway Plugins](https://aifilmhq.com/tool/runway-plugins/): 1 entry - [SolarRenderer](https://aifilmhq.com/tool/solarrenderer/): 1 entry - [Sync Lipsync v3](https://aifilmhq.com/tool/sync-lipsync-v3/): 1 entry - [Vidu Q4](https://aifilmhq.com/tool/vidu-q4/): 1 entry - [Whistle](https://aifilmhq.com/tool/whistle/): 1 entry ## New - [New by day](https://aifilmhq.com/new/): 0 new today, 25 new this week ## Learn - [Learn](https://aifilmhq.com/learn/): tutorials, guides, videos and courses (48 items) - [Editing](https://aifilmhq.com/learn/editing/): 4 resources - [AI film pipeline](https://aifilmhq.com/learn/ai-film-pipeline/): 6 resources - [Upscaling pipelines](https://aifilmhq.com/learn/upscaling-pipelines/): 2 resources - [ComfyUI video workflows](https://aifilmhq.com/learn/comfyui-video-workflows/): 7 resources - [Lip sync](https://aifilmhq.com/learn/lip-sync/): 1 resource - [Consistent characters](https://aifilmhq.com/learn/consistent-characters/): 6 resources - [Sound for AI films](https://aifilmhq.com/learn/sound-for-ai-films/): 1 resource - [Shot prompting](https://aifilmhq.com/learn/shot-prompting/): 4 resources - [Pre-production](https://aifilmhq.com/learn/pre-production/): 5 resources - [Sound](https://aifilmhq.com/learn/sound/): 2 resources - [Lenses](https://aifilmhq.com/learn/lenses/): 2 resources - [Framing and composition](https://aifilmhq.com/learn/framing-and-composition/): 3 resources - [Camera movement](https://aifilmhq.com/learn/camera-movement/): 1 resource - [Storytelling](https://aifilmhq.com/learn/storytelling/): 2 resources - [Lighting](https://aifilmhq.com/learn/lighting/): 1 resource - [Color grading](https://aifilmhq.com/learn/color-grading/): 1 resource - [Google Flow Tutorial: Transitions](https://aifilmhq.com/learn/google-flow-transitions-tutorial/): This weekly Google Flow tutorial by Jerrod Lew focuses on transitions between shots. It shows how to join generated clips so the cut feels intended. - [Script to Storyboard: Plan Your AI Short Film](https://aifilmhq.com/learn/script-to-storyboard-ai-short-film/): Vixel's guide turns a short script into a shot list and then a storyboard using worked examples. Its main lesson is to turn every feeling in the script into an action the camera can see. - [LTX-2.5 DFR in ComfyUI: Image to Video with Audio](https://aifilmhq.com/learn/ltx-2-5-dfr-comfyui/): A three-stage LTX-2.5 pipeline that generates video with synced audio from one image. It renders at half size, upscales 2x spatially, then optionally doubles the frame rate. - [You Already Know LTX, Part 4: Make Your Own IC-LoRA](https://aifilmhq.com/learn/you-already-know-ltx-part-4-train-ic-lora/): In this long X article Christopher Fryant explains, from zero, how he trained the Blackout IC-LoRA for LTX. It covers what LoRAs and IC-LoRAs are, how to build before-and-after training pairs and how trigger words work. - [LTX-Video in ComfyUI: T2V, I2V, Multi-Frame Control](https://aifilmhq.com/learn/comfyui-ltx-video-workflows/): ComfyUI's guide to LTX-Video text-to-video, image-to-video and multi-frame control workflows. It stresses that this model needs long, descriptive prompts. - [Wan2.2-S2V in ComfyUI: Audio-Driven Video](https://aifilmhq.com/learn/comfyui-wan-2-2-s2v-lip-sync/): ComfyUI's native workflow for Wan2.2-S2V, which turns a still image plus audio into a talking or singing video. It supports full-body and half-body characters and long-form output. - [AI B-Roll Workflow: Pick Moments From the Script, Direct a Frame, Animate in Runway](https://aifilmhq.com/learn/ai-b-roll-workflow-chatgpt-reference-frames-runway/): Matt Paige shows how he makes B-roll for short videos: ChatGPT picks the three or four script lines that need a visual, he builds and fixes a reference frame, then animates it with Runway inside ChatGPT. The tutorial gives his prompts for each step and includes a Runway discount code. - [You Already Know LTX, Part 3: The Blackout Flashlight Effect](https://aifilmhq.com/learn/you-already-know-ltx-part-3-blackout/): Christopher Fryant shares his first own LTX IC-LoRA, Blackout, which turns the lights off in a scene and relights it with a flashlight. The thread includes the steps to get the effect running in ComfyUI. - [You Already Know LTX, Part 2: Add Water to Any Scene](https://aifilmhq.com/learn/you-already-know-ltx-part-2-water/): Part 2 of Christopher Fryant's series applies an LTX water-simulation IC-LoRA to existing footage. It adds rain, wet surfaces, puddles and splashes, and the prompt can steer the result a little. - [Qwen-Image-Edit-2511 in ComfyUI: Multi-Reference Editing](https://aifilmhq.com/learn/comfyui-qwen-image-edit-2511/): The official ComfyUI docs show a local workflow for Qwen-Image-Edit-2511, an open image-editing model. It keeps a character's identity across edits and can combine several reference images, which helps build character and location plates locally. - [Reference Sheets Are a Cheat Code for AI Video](https://aifilmhq.com/learn/reference-sheets-for-ai-video/): Jerrod Lew shows how to make detailed reference sheets for characters, objects and environments with image models. Feeding these sheets to video tools gives them more context and keeps generations consistent. - [Google Flow Tutorial: Storyboards](https://aifilmhq.com/learn/google-flow-storyboards-tutorial/): Jerrod Lew's tutorial, made with Google Flow, covers building a storyboard inside Flow. He shows how the Flow agent helps plan shots and keep the video consistent. - [You Already Know LTX, Part 1: Remove People With a Clean-Plate IC-LoRA](https://aifilmhq.com/learn/you-already-know-ltx-part-1-clean-plate/): Christopher Fryant starts an X series on running custom LTX video workflows in ComfyUI without specialist knowledge, letting an AI agent handle the setup. Part 1 uses a clean-plate IC-LoRA that removes people from any video. - [Sound Design for AI Short Films: A Five-Pass Workflow](https://aifilmhq.com/learn/sound-design-ai-short-films-five-pass/): Prompt2Screen gives a five-pass order for adding sound to an AI short once the picture is locked. It builds from a spotting list and ambience beds to spot effects, music and the final mix. - [ComfyUI Image Upscale Workflow](https://aifilmhq.com/learn/comfyui-image-upscale/): A beginner tutorial on AI upscaling in ComfyUI with ESRGAN models. It shows how to download models from OpenModelDB and chain upscaling after text-to-image. - [Wan2.2 Video Generation in ComfyUI](https://aifilmhq.com/learn/comfyui-wan-2-2-workflow/): The official ComfyUI tutorial for running Wan 2.2 video models with native workflows. It covers the model versions, downloads and the 5B hybrid text/image-to-video workflow. - [AI Camera Prompts: Movement, Angles and Shots in AI Video](https://aifilmhq.com/learn/ai-camera-prompts-guide/): Runway's guide lists the camera terms video models follow and gives an order for putting them into a prompt. It explains why named moves like a slow dolly push-in work better than words like cinematic. - [AI Shot List: How to Build a Camera-Ready Plan](https://aifilmhq.com/learn/ai-shot-list-camera-ready-plan/): Runway explains how to turn a script, scene or brief into a shot list with framing, angle, movement and action for every shot. It also shows how to write shot descriptions that can be used directly as video prompts. - [LTX-2.5 Image-to-Video and Text-to-Video Workflow Guide](https://aifilmhq.com/learn/ltx-2-5-workflow-guide/): LTX's own walkthrough of the official ComfyUI workflows for LTX-2.5. It explains multiscale rendering, audio-video generation and when to use the distilled or full model. - [Wan2.2 Animate: Character Animation and Replacement](https://aifilmhq.com/learn/comfyui-wan-2-2-animate/): The official ComfyUI workflow for Wan2.2 Animate. It animates a character from a performer's video or replaces a person in a video while keeping lighting and color. - [What is a Shot List?](https://aifilmhq.com/learn/what-is-a-shot-list/): This guide shows what goes on a shot list and how to make one in five steps. It also explains how a shot list differs from a storyboard. - [AI Short Film Workflow: From Generated Clips to a Finished Edit](https://aifilmhq.com/learn/ai-short-film-workflow-davinci-resolve/): Envato walks through making a cinematic short with an AI video generator and finishing it in DaVinci Resolve. It covers planning the shots, generating the clips and assembling them into one film. - [What is Diegetic Sound?](https://aifilmhq.com/learn/diegetic-sound-explained/): This guide explains the difference between diegetic and non-diegetic sound. It covers diegetic dialogue, music and sound effects with film examples. - [What Is Room Tone in Film?](https://aifilmhq.com/learn/room-tone-guide/): A practical guide to recording room tone and using it in the edit. It includes a 7-step on-set method and common mistakes to avoid. - [What is Depth of Field? Shallow vs Deep](https://aifilmhq.com/learn/depth-of-field-explained/): This guide defines depth of field and contrasts shallow and deep focus. It explains what affects depth of field and how to adjust it with aperture. - [How to Write a Creative Brief for Video Production](https://aifilmhq.com/learn/creative-brief-video-production/): Celtx explains what a creative brief is and the seven parts every video brief should have. It then walks through writing one so everyone agrees on the goal before the script. - [50+ Types of Camera Shots, Angles, and Techniques](https://aifilmhq.com/learn/types-of-camera-shots/): A reference guide to every common shot size, angle and framing technique. It runs from extreme wide shots to extreme close-ups with film examples for each. - [Red Rising Teaser: Book Passage to Trailer With Luma's Agent and Seedance 2.0](https://aifilmhq.com/learn/red-rising-teaser-luma-agent-seedance-workflow/): PJ Ace breaks down how he turned a 20-minute audiobook scene from Red Rising into a 3-minute fan teaser in under a week, with prompts at each step. He used Luma's agent for scripting help, character looks and 2x2 image grids, then animated the chosen frames in Seedance 2.0; the thread was posted on Luma agent's launch day and promotes it. - [Kling VIDEO 3.0 Model Guide](https://aifilmhq.com/learn/kling-video-3-model-guide/): Kling's official guide to its VIDEO 3.0 model. It covers multi-shot narratives, image-to-video with subject consistency, native audio and 15-second generation. - [Kling VIDEO 3.0 Omni Guide](https://aifilmhq.com/learn/kling-video-3-omni-guide/): Kling's guide to the 3.0 Omni model for reference-driven video. It explains all-in-one references, Elements 3.0 character capture and storyboard narration. - [Consistent Character AI Video: Two-Frame Tutorial](https://aifilmhq.com/learn/two-frame-consistent-character/): A six-step workflow that turns one approved portrait into two keyframes and a short consistent clip. It includes prompts, a frame-by-frame acceptance sheet and drift troubleshooting. - [The Rule of Thirds in Film](https://aifilmhq.com/learn/rule-of-thirds-guide/): StudioBinder explains the rule of thirds and how filmmakers place subjects on a 3x3 grid. It compares the rule with centered framing and shows how to plan it in a shot list. - [A Six-Layer Framework for Prompting AI Video Models](https://aifilmhq.com/learn/six-layer-video-prompt-framework/): Venice AI breaks a video prompt into six layers: subject and action, shot type, camera movement, lighting, technical specs and pacing. It also shares three ways to iterate that save credits across models like Kling, Sora, Wan and Veo. - [Ultimate Prompting Guide for Veo 3.1](https://aifilmhq.com/learn/veo-3-1-prompting-guide/): Google Cloud's guide to writing prompts for Veo 3.1 video generation. It gives a prompt formula, core techniques and advanced creative workflows. - [How I Created an Animated Short Film Using AI in 8 Steps](https://aifilmhq.com/learn/animated-short-film-ai-8-steps/): Shweta Kaushal lays out an eight-step pipeline from a story idea to a finished animated short. Each step names the kind of tool used, from language models for the script to diffusion models and AI voice. - [How I Made My First AI-Native Film: The Old Man & The Bee](https://aifilmhq.com/learn/first-ai-native-film-old-man-and-the-bee/): db clay describes making a personal short film in about a day for under $100 with generative tools. The post compares each of the five steps with a traditional film workflow. - [Definitive Guide to Every Type of Camera Movement in Film](https://aifilmhq.com/learn/camera-movement-guide/): This guide covers each type of camera movement and when to use it. It starts with the static frame and moves through pans, tilts and more complex moves. - [Three Act Structure in Film](https://aifilmhq.com/learn/three-act-structure/): StudioBinder explains the three-act structure, its history and why it works. It also covers why some screenwriters disagree with it. - [The Kuleshov Effect Explained](https://aifilmhq.com/learn/kuleshov-effect-explained/): StudioBinder explains the Kuleshov effect, where meaning comes from how shots are put side by side. It shows modern examples and how Spielberg uses reaction shots. - [Three-Point Video Lighting: Key, Fill and Backlight](https://aifilmhq.com/learn/three-point-lighting-setup/): A step-by-step guide to the classic three-point lighting setup. It defines the key, fill and backlight and shows how to build each one. - [Types of Editing Transitions in Film](https://aifilmhq.com/learn/editing-transitions-guide/): A guide to the cuts and transitions editors use to move between shots. It covers each type with film examples. - [Gen-4 References: Consistent Characters and Locations](https://aifilmhq.com/learn/runway-gen-4-references-tutorial/): This Runway Academy video shows how to keep a character and a location consistent by feeding up to three reference images. Saved references can be called by name in later prompts. - [How to Make a Storyboard](https://aifilmhq.com/learn/how-to-make-a-storyboard/): StudioBinder explains how to plan a film frame by frame with a storyboard. It covers what goes into each panel and how a board connects to the script and shot list. - [Focal Length: Understanding Camera Lenses](https://aifilmhq.com/learn/focal-length-lenses-explained/): StudioBinder explains what focal length is and why it changes the look of a shot. It compares wide-angle, standard and long-focus lenses. - [What Is the 180 Degree Rule in Film and How to Break It](https://aifilmhq.com/learn/180-degree-rule-explained/): StudioBinder explains the imaginary line that keeps characters' positions and eyelines consistent between cuts. It also shows when and how filmmakers cross it on purpose. - [What Is a Creative Brief: Types, Templates and Formats](https://aifilmhq.com/learn/what-is-a-creative-brief/): StudioBinder defines the creative brief and the elements it should contain. It covers objectives, audience, key messages, deliverables, timeline and budget, with tips for keeping it short. - [What Are the Basics of Color Grading?](https://aifilmhq.com/learn/color-grading-basics/): No Film School explains the difference between colour correction and colour grading and the basic steps of each. It is a starting point for giving a set of shots one consistent look. - [How to Format a Screenplay](https://aifilmhq.com/learn/how-to-format-a-screenplay/): Sam Kench at StudioBinder covers standard screenplay format, from scene headings to action lines and dialogue. A properly formatted script is also easier to break into shots later. ## Questions and answers ### https://aifilmhq.com/category/video-generation/ - Q: Which AI video generators can I run locally for free? A: Among the entries here, MiniMax H3 and Kandinsky 6.0 Video have open weights you can run on your own GPU. WanGP and ComfyUI are free apps that run these models locally. - Q: Which AI video models generate sound with the video? A: Several entries here make audio in the same pass, including MiniMax H3 (stereo sound), Kandinsky 6.0 Video and HeyGen Video. Kandinsky's clips are limited to 5 seconds. - Q: How much does AI video generation cost per second? A: Prices in our entries vary: Boreal is $0.01 per second, H3 Max is $0.08 per second at 768p and BACH 1.0 Pro is listed at $0.12 per second for 1080p. Google Vids offers free AI video with a Google account. - Q: Which AI video generator keeps characters consistent across shots? A: Among the entries here, Vidu Q4 accepts up to 15 reference images and 3 audio clips, and BACH 1.0 Pro takes up to 12 references per generation. See also the consistent characters guides. ### https://aifilmhq.com/category/image/ - Q: What is a good open-weight image model I can run on 8 GB of VRAM? A: Among the entries here, FLUX 3 Image Klein has an INT8 build that runs on GeForce RTX GPUs with as little as 8 GB VRAM via ComfyUI or Diffusers. It uses the FLUX Kommunity License, which is not fully permissive. - Q: Which AI image tools are new this autumn? A: Recent entries include ChatGPT Images 2.5, Krea 3, MAI-Image-2.6 and FLUX 3 Image Klein. ComfyUI added FLUX.3 Image support in v0.41.0. - Q: How much does Krea 3 cost per image? A: Krea 3 costs $0.06 per image at 2K. It is live on krea.ai and through several API providers such as fal and Replicate. ### https://aifilmhq.com/category/voice-and-audio/ - Q: Which text-to-speech models can I run locally? A: Among the entries here, Fish Audio Drama 3 has weights on Hugging Face under the Fish Audio Community License. For speech-to-text, Whistle is a 16.9 MB model that runs on a CPU. - Q: What is the cheapest TTS API for video voiceover? A: Prices in our entries include Nari Qwen3-TTS from $5 per 1M characters, Fish Audio Drama 3 at $10 per 1M until November 12, and MAI-Voice-2.1 Flash at $15 per 1M. Quality and language support differ, so test with your own script. - Q: How can I make captions for my video with AI offline? A: Whistle transcribes seven languages with word timestamps on your own device, so audio never leaves it. Longer audio is processed in windows. - Q: Can I clone a voice with Gemini TTS in Europe? A: No, according to our Gemini 3.8 Flash TTS entry, voice replication isn't available in the EEA, the UK, Switzerland, India, Illinois or Texas. Voice design and the other features are available in Google AI Studio and the Gemini API. - Q: How can I add music and sound effects to a finished video with AI? A: ElevenLabs Instant Soundtracks scores an uploaded video and places timed sound effects on separate layers you can edit. Mirelo SFX 2.0 keeps the dialogue and rebuilds effects, foley and ambience for edits up to 10 minutes. To write and render songs locally, Plenio runs inside ComfyUI. ### https://aifilmhq.com/category/editing/ - Q: Is there a free open-source video editor with AI masking? A: OpenShot 4.0 is free and open source and can isolate subjects with machine learning models that run locally. It also adds color wheels, curves, LUTs and scopes. - Q: Can I use AI video generation inside Premiere Pro? A: Yes: our entries cover Adobe's Generative Media Tool in the Premiere timeline and the free Runway plugins for Premiere Pro and After Effects. Generating with Runway needs a paid Runway plan. - Q: How do I join many AI video clips together in ComfyUI? A: ComfyUI v0.36.0 added a Video Concatenate node that joins up to 100 clips in one graph. You need v0.36.0 or newer. - Q: How do I remove music an AI video model added to my clip? A: Mirelo SFX 2.0 detects and keeps the dialogue, removes the other audio, such as unprompted music, and generates new effects and ambience synced to the action, so you don't have to regenerate the video. It's available on Mirelo's own tools, Runway and fal. - Q: Is there a free, local alternative to CapCut? A: Concat is a free, open-source editor for macOS, Windows, Linux and Android with auto-captions, text-to-speech, voice cloning and background removal that run on your machine. It's still in beta and has no watermark or account. ### https://aifilmhq.com/category/upscaling/ - Q: How can I upscale AI video locally? A: Among the entries here, NVIDIA VFX Nodes run inside ComfyUI on an RTX GPU and upscale video up to 4×. They also do frame generation and SDR-to-HDR conversion. - Q: What order should I run upscaling, frame interpolation and HDR? A: Comfy's announcement for NVIDIA VFX Nodes recommends Super Resolution, then Frame Generation, then HDR. The upscaling pipelines guides show other ComfyUI chains. - Q: Can I upscale video with Runway in Premiere? A: The Runway plugins for Premiere Pro and After Effects are tagged for upscaling here. The plugins are free, but generating uses paid Runway plan credits. ### https://aifilmhq.com/category/lip-sync/ - Q: What is the best local AI lip sync tool? A: Among the entries here, Kandinsky 6.0 Video is the open-weight option: it generates lip-synced speech with the video under the MIT license. For audio-driven talking video in ComfyUI, see the Wan2.2-S2V guide. - Q: Which AI lip sync tool is good for dubbing and localization? A: LipDub 2.0 focuses on short-form localization with no training fees, and Sync Lipsync v3 is available on Artlist. Sync Lipsync v3 needs an audio file and has no text prompt. - Q: Does lip sync need an audio file? A: For Sync Lipsync v3, yes: audio is required and outputs are MP4. Audio-video models like Kandinsky 6.0 Video instead generate the speech themselves. - Q: Can I make a talking-head video with open models on my own GPU? A: Yes: the free drama3-h3-talking ComfyUI workflow turns text into speech with Fish Audio Drama 3, then drives a lip-synced MiniMax H3 video from a face image. Its author measured about 130 to 190 seconds for the video stage of a 7 to 9 second clip on a 96 GB card, and both models have community licenses to check first. ### https://aifilmhq.com/category/workflows/ - Q: What is ComfyUI used for in AI filmmaking? A: ComfyUI is a free, open-source node-based app for building image, video and audio workflows that run locally. Our entries cover its releases, such as v0.40 and v0.41. - Q: Can I turn a Blender or Unreal blockout into an AI video? A: LTX-2.5 Layout to Render is an IC-LoRA that turns a grey viewport or playblast into a finished shot, with an art-directed first frame setting the look. Both the layout video and the first frame are required. - Q: Has anyone made a short film with local AI on a home PC? A: Yes: Chris Beach made a short space film with MiniMax H3 in ComfyUI on one RTX 5090. It took about 32 hours of rendering and roughly £5.44 in electricity. ### https://aifilmhq.com/category/local/ - Q: What do I need to run AI video locally? A: Our entries show a range: MiniMax H3 runs on a 12 GB RTX 3060 with plenty of system RAM, and WanGP v17 reports 480p in 5-6 GB VRAM. On a Mac, h3-metal runs H3 on Apple Silicon. - Q: Which apps let me run open AI video models on my own PC? A: Among the entries here, ComfyUI and WanGP are the main free local apps. OpenShot 4.0 is a free local editor with on-device AI masks. - Q: Can I run AI video on a Mac? A: h3-metal is an open-source C and Metal engine that runs MiniMax H3 on Apple Silicon. Its README reports renders of about 75 seconds on a 128 GB M5 Max but gives no minimum spec. - Q: Is there an AI film studio app that runs on my own GPU? A: Mago Local is a Windows app that takes an idea or script to a finished film, or restyles your footage, on an NVIDIA GPU with 16 GB+ VRAM, and local renders don't use credits. WanGP v18.17 adds a Deepy Prime agent that plans a one-minute video from a brief. ### https://aifilmhq.com/tool/comfyui/ - Q: Is ComfyUI free? A: Yes, ComfyUI is free and open source and runs on your own computer. Individual models keep their own licenses; for example, commercial use of local MiniMax H3 outputs is licensed separately through Comfy. - Q: What's new in the latest ComfyUI? A: v0.41.0 adds FLUX.3 Image and MiniMax H3 LightVAE support, v0.40.0 adds camera angle nodes and LoRA stack loaders, and v0.36.0 adds Video Concatenate and FastH3. - Q: How do I update ComfyUI? A: Our v0.36 entry says to update through the Manager or your launcher. New workflow templates often need a minimum version, so check it first. - Q: Where do I start with video workflows in ComfyUI? A: The ComfyUI video workflows learn page collects official guides for Wan 2.2 and LTX. For MiniMax H3, start with the H3 in ComfyUI entry. ### https://aifilmhq.com/tool/minimax-h3/ - Q: Can I run MiniMax H3 on a 12 GB GPU? A: Yes, ComfyUI's pruned, quantized files can run MiniMax H3 on a 12 GB RTX 3060. A cited owner run took about 4.5 minutes for 5 seconds at 864×480 with an 8-step Turbo LoRA, and host RAM matters: plan on 32 GB or more. - Q: Do I need a license to use MiniMax H3 in the EU? A: Users in the EU, UK, US and South Korea request a license from MiniMax through its H3 License Request Form before use; the link is in our H3 entry. Commercial use of local outputs is licensed separately through Comfy. - Q: How long can MiniMax H3 videos be? A: MiniMax H3 makes clips with stereo sound up to 2K and 15 seconds. WanGP v17 reports 15 seconds at 1080p in 11 GB VRAM. - Q: Is there a faster version of MiniMax H3? A: Yes: ComfyUI ships the 8-step FastH3, and FastH3-Omni is in preview with open weights announced for October 20, 2026. A hosted H3 Max is also on fal. - Q: Can MiniMax H3 follow a camera move I block out in Blender? A: Yes, roughly: in this Blender blockout use case, H3's reference-to-video mode filmed a gray Blender flight path as a 22-second drone shot on one RTX 5090. It held the order of landmarks and the direction of travel, but softened exact moves like a full orbit. - Q: Can MiniMax H3 turn a grey-box or skeleton video into a finished shot? A: Yes, with the open SolarRenderer LoRAs: they render a structure video, such as a grey box, skeleton, depth pass or pencil storyboard, into finished video in four steps, with the look set by the first frame and the prompt. It needs the full H3 weights (about 135 GB) and runs from Python scripts. ### https://aifilmhq.com/tool/wangp/ - Q: What is WanGP? A: WanGP is a free local app for open video, image and audio models. Recent versions added MiniMax H3 memory savings in v17 and Kandinsky 6.0 in v18. - Q: How much VRAM does WanGP need for MiniMax H3? A: The developer says WanGP v17 needs 11 GB for 15 seconds of H3 at 1080p and 5-6 GB at 480p. Results depend on your profile, attention backend and system RAM. - Q: Why is my first WanGP generation slow on an RTX 50 series card? A: Our v18 entry says to tick Config > Performance > WanGP Compute Cache to avoid a 10-15 second delay at each first generation. - Q: Can WanGP make dialogue with cloned voices? A: Yes: WanGP v18.17 adds Fish Audio Drama 3, which writes speaker lines with emotions, whispers, laughs and pauses and clones up to three voices. An INT4 Experts variant roughly halves the experts' memory. ### https://aifilmhq.com/tool/flux-3-image-klein/ - Q: Can FLUX 3 Klein run on 8 GB VRAM? A: Yes, the INT8 build of FLUX 3 Image Klein runs on GeForce RTX GPUs with as little as 8 GB VRAM via ComfyUI or Diffusers. - Q: Is FLUX 3 Klein free for commercial use? A: It's released under the FLUX Kommunity License, which is not a fully permissive license, so read its terms before commercial use. It's also available in the BFL API. - Q: Does ComfyUI support FLUX 3 Image? A: Yes, ComfyUI v0.41.0 added native support for FLUX.3 Image models, including Klein. - Q: Can I run FLUX 3 Klein on a Mac? A: Yes: mflux 0.23 runs FLUX 3 Image Klein natively on Apple Silicon with MLX, including editing with up to ten reference images, LoRAs and 4, 6 or 8-bit quantization. You accept the FLUX Kommunity License on Hugging Face before the first download. ### https://aifilmhq.com/tool/kandinsky-6-0-video/ - Q: Is Kandinsky 6.0 Video open source? A: Yes, Kandinsky 6.0 Video code and weights are released under the MIT license, with Lite and Pro models. - Q: How much VRAM does Kandinsky 6.0 need? A: The paper includes memory presets for 32, 24 and 16 GB GPUs. It's also available in WanGP v18. - Q: How long are Kandinsky 6.0 videos? A: Clips top out at 5 seconds, base generation is at SD resolution and Full HD comes from upscaling, according to our entry. ### https://aifilmhq.com/tool/fasth3-omni/ - Q: When are FastH3-Omni weights released? A: Open weights and training code for FastH3-Omni are announced for October 20, 2026. Until then it's a hosted preview. - Q: Does FastH3 need the H3 license? A: FastH3 uses the H3 weights, so the same H3 license form applies; see our ComfyUI v0.36 entry for the link. - Q: How fast is FastH3? A: No measured time for FastH3-Omni is public yet. The earlier FastH3 V2 took 36.5 seconds for a 5-second 1344×768 clip on one RTX PRO 6000. ### https://aifilmhq.com/learn/shot-prompting/ - Q: How do I write a good prompt for AI video? A: The Veo 3.1 prompting guide gives a prompt formula and core techniques, and the Kling VIDEO 3.0 guide covers multi-shot prompts. Knowing shot types and camera moves helps you name what you want. - Q: Do AI video prompts need to be long? A: For LTX-Video, yes: ComfyUI's LTX guide stresses that the model needs long, descriptive prompts. - Q: Can I prompt multiple shots in one AI video? A: The Kling VIDEO 3.0 guide covers multi-shot narratives with native audio and 15-second generation. ### https://aifilmhq.com/learn/consistent-characters/ - Q: How do I keep the same character across AI video shots? A: The two-frame tutorial turns one approved portrait into two keyframes and a short clip, with drift troubleshooting. Reference-driven models like Kling 3.0 Omni and Vidu Q4 also take character references. - Q: Can I replace an actor with an AI character in ComfyUI? A: Yes, the Wan2.2 Animate workflow animates a character from a performer's video or replaces a person while keeping lighting and color. - Q: How many reference images can AI video models use? A: Among our entries, Vidu Q4 takes up to 15 images and 3 audio clips, and BACH 1.0 Pro up to 12 references. ### https://aifilmhq.com/learn/comfyui-video-workflows/ - Q: How do I make AI video in ComfyUI? A: Start with the official Wan2.2 tutorial or the LTX-Video guide, which cover text-to-video and image-to-video workflows. The LTX-2.5 workflow guide adds audio-video generation. - Q: Should I use the distilled or full LTX-2.5 model? A: The LTX-2.5 workflow guide explains when to use each one in the official ComfyUI workflows. - Q: Which video models work in ComfyUI? A: Our guides and entries cover Wan 2.2, LTX-2.5, MiniMax H3 and FastH3, among others. See the ComfyUI tool page for release entries. ### https://aifilmhq.com/learn/upscaling-pipelines/ - Q: How do I upscale images in ComfyUI? A: The ComfyUI Image Upscale tutorial shows how to download ESRGAN models from OpenModelDB and chain upscaling after text-to-image. - Q: How do I upscale AI video in ComfyUI? A: The LTX-2.5 DFR pipeline renders at half size, upscales 2x and can double the frame rate. NVIDIA VFX Nodes upscale video up to 4× on RTX GPUs. - Q: Is it faster to render small and upscale AI video? A: The LTX-2.5 DFR guide uses that approach, rendering at half size before a 2x spatial upscale. Kandinsky 6.0 also reaches Full HD by upscaling from SD. ### https://aifilmhq.com/learn/lip-sync/ - Q: How do I make a talking video from a photo and audio? A: ComfyUI's Wan2.2-S2V workflow turns a still image plus audio into a talking or singing video, with full-body and half-body support. - Q: Can AI lip sync make long videos? A: The Wan2.2-S2V guide says it supports long-form output. Hosted tools like LipDub 2.0 focus on dubbing existing footage. - Q: Which lip sync tools are in this directory? A: See the lip sync category, which includes Sync Lipsync v3, LipDub 2.0 and Kandinsky 6.0 Video. ## Entries - [Use case: a music video made fully locally with MiniMax H3, YuE2 and HiDream O1](https://aifilmhq.com/entry/use-case-music-video-made-fully-locally-with-minimax-h3-yue2-and-hidream-o1/): A creator made a three-minute music video using only open-weight models on their own machine: YuE2 for the song, HiDream O1 for character and world references, Qwen 3.8 27B to format prompts and MiniMax H3 for the video. The edit was done in DaVinci Resolve with only basic tools and no cloud AI. - [Kling 4.0 launches globally](https://aifilmhq.com/entry/kling-4-0-launches-globally/): Kling 4.0 is Kling AI's new video model, which can generate up to 30 seconds of video in one go with keyframe control and sound. After early access in late September, it is now fully available, including through MCP and a command-line tool. - [mflux 0.23 runs FLUX 3 Image Klein locally on Apple Silicon](https://aifilmhq.com/entry/mflux-0-23-runs-flux-3-image-klein-on-apple-silicon/): The open-source mflux tool now runs Black Forest Labs' FLUX 3 Image Klein natively on Macs with MLX. Version 0.23 covers text-to-image, editing with up to ten reference images, LoRAs and 4, 6 or 8-bit quantization. - [Open ComfyUI workflow turns text into a lip-synced talking video with Drama 3 and MiniMax H3](https://aifilmhq.com/entry/open-comfyui-workflow-drama-3-voice-to-minimax-h3-talking-video/): A free ComfyUI workflow from Skysay chains Fish Audio Drama 3 for expressive speech with MiniMax H3 for a lip-synced talking video, all on your own GPU. It ships as MIT-licensed graphs plus a script that runs both stages. - [OpenMOSS releases MOSS-Image, an open text-to-image model with layout control](https://aifilmhq.com/entry/openmoss-releases-moss-image-open-text-to-image-model/): MOSS-Image is a new open-weight text-to-image family from OpenMOSS, with Base and SFT checkpoints under Apache 2.0. Besides plain prompts, it accepts tag lists and JSON layouts that place text and objects with bounding boxes. - [Pika's ChatGPT plugin puts its video, image and sound apps inside the chat](https://aifilmhq.com/entry/pika-plugin-brings-pika-video-image-and-sound-apps-into-chatgpt/): Pika launched a plugin for ChatGPT that runs its 40+ creation apps and 55+ image, video, music and speech models inside the chat. You can ask ChatGPT to bring up an editing timeline, and the plugin works with your existing Pika library. - [Plenio 1.0 turns ComfyUI into a free local song studio](https://aifilmhq.com/entry/plenio-1-0-turns-comfyui-into-a-local-song-studio/): Plenio is a free, open-source ComfyUI node pack that goes from a song idea to a mastered track on your own machine. A local language model writes the lyrics, YuE2 or MiniMax Music 3 performs them, and version 1.0 adds a browser player and lyric videos. - [Use case: a 22-second drone flight over Taiwan from a Blender blockout and MiniMax H3](https://aifilmhq.com/entry/use-case-blender-blockout-drone-flight-with-minimax-h3-on-one-rtx-5090/): A creator made a 22-second FPV-style flight past four Taiwanese landmarks without a drone, using a gray Blender blockout and MiniMax H3 running locally on one RTX 5090. The write-up shares every step, the render times and what still breaks. - [WanGP v18.17 adds FLUX 3 Klein, Fish Audio Drama 3 and MOSS Image](https://aifilmhq.com/entry/wangp-v18-17-adds-flux-3-klein-drama-3-and-moss-image/): WanGP v18.17 brings FLUX 3 Image Klein, Fish Audio Drama 3 dialogue with voice cloning and MOSS Image 5B to the low-VRAM local app. It also adds a lighter H3 video decoder, INT8 math kernels and a Deepy Prime agent that plans a one-minute video from a brief. - [Black Forest Labs opens FLUX 3 Image Klein](https://aifilmhq.com/entry/black-forest-labs-opens-flux-3-image-klein/): FLUX 3 Image Klein is a 5.6B-parameter open-weights model for image generation, editing and multi-reference work with up to 10 images. It supports native 4K and bounding boxes to control where things go. - [ComfyUI v0.41.0 adds FLUX.3 Image and MiniMax H3 LightVAE support](https://aifilmhq.com/entry/comfyui-v0-41-flux-3-image-and-h3-lightvae/): ComfyUI v0.41.0 adds native support for Black Forest Labs' FLUX.3 Image models, including the open FLUX 3 Image Klein, and for MiniMax H3's LightVAE. It also adds ImageSDRToHDR and ReferenceMultiLatent nodes. - [ElevenLabs Instant Soundtracks scores a video and places sound effects in one pass](https://aifilmhq.com/entry/elevenlabs-instant-soundtracks-scores-video-with-timed-sound-effects/): Instant Soundtracks in ElevenCreative takes an uploaded video, writes an original score for it and places sound effects at the moments that call for them. The effects land on their own layers, and the API can return music and effects as separate files. - [Krea 3 image editing model](https://aifilmhq.com/entry/krea-3-image-editing-model/): Krea 3 is an image editing model that returns most edits in under 8 seconds at native 2K while keeping products and characters consistent. It costs $0.06 per image and is also offered via partners like fal and Replicate. - [Mago Local runs the Mago AI film studio on your own GPU](https://aifilmhq.com/entry/mago-local-runs-the-mago-ai-film-studio-on-your-own-gpu/): Mago released a Windows desktop version of its AI video studio that renders on your own NVIDIA GPU, so local renders don't use credits. It can connect to an existing ComfyUI install or set one up, and the cloud models are still available in the app. - [Sarvam launches Bulbul V4 Flash for Indian languages](https://aifilmhq.com/entry/sarvam-launches-bulbul-v4-flash-for-indian-languages/): Bulbul V4 Flash is a fast text-to-speech model with 83 voices across 11 Indian languages. It reaches first audio in under 250 ms and costs ₹20 per 10,000 characters. - [SolarRenderer turns grey-box, skeleton or sketch video into finished shots with MiniMax H3](https://aifilmhq.com/entry/solarrenderer-turns-grey-box-skeleton-or-sketch-video-into-finished-shots-with-minimax-h3/): SolarRenderer is a pair of open LoRAs for MiniMax H3 that renders a structure video, such as a game-engine grey box, a skeleton or depth pass, or a pencil storyboard, into finished video in four sampling steps. The first frame and the prompt set the look, so the same blockout can come out photoreal, anime, pixel art or ink. - [Video Rebirth launches BACH 1.0 Pro](https://aifilmhq.com/entry/video-rebirth-launches-bach-1-0-pro/): BACH 1.0 Pro is a video model aimed at film, advertising and animation teams that need consistent characters and products across shots. Each generation can use up to 12 image, video or audio references, and 1080p API output is listed at $0.12 per second. - [FastH3-Omni preview speeds up MiniMax H3 reference video](https://aifilmhq.com/entry/fasth3-omni-preview-speeds-up-minimax-h3-reference-video/): FastH3-Omni is an accelerated version of MiniMax H3 that conditions clips on images, audio, text and first and last frames. A hosted preview is live, with open weights and training code announced for October 20, 2026. - [Gnani Timbre v3.0 adds audio tags](https://aifilmhq.com/entry/gnani-timbre-v3-0-adds-audio-tags/): Timbre v3.0 is an Indian-language text-to-speech model that uses inline tags to make a voice laugh, sigh, whisper or pause. It launches with 13 voices in 10 languages. - [Mirelo SFX 2.0 rebuilds the sound of a video and keeps the dialogue](https://aifilmhq.com/entry/mirelo-sfx-2-0-rebuilds-the-sound-of-a-video-and-keeps-the-dialogue/): Mirelo SFX 2.0 is a video-to-sound model that removes unwanted music from a clip, keeps the dialogue, and generates effects, foley and ambience synced to the action. It handles edits of up to 10 minutes in one pass and is now also available on Runway. - [ComfyUI v0.40.0 adds camera-angle nodes and LoRA stack loaders](https://aifilmhq.com/entry/comfyui-v0-40-camera-angle-nodes-and-lora-stacks/): ComfyUI v0.40.0 adds CameraAngle and CreateCameraInfo nodes for defining camera parameters inside a workflow, plus native LoRA stack loaders. It also changes MiniMax H3 image decoding and fixes EXR and half-float file handling. - [Concat is a free, open-source CapCut alternative that runs its AI locally](https://aifilmhq.com/entry/concat-free-open-source-capcut-alternative-with-local-ai/): Concat is an open-source video editor for macOS, Windows, Linux and Android with auto-captions, text-to-speech with voice cloning and background removal that run on your own machine. It's in beta, has no watermark or account, and scripts or AI agents can drive it through a JSON-RPC API. - [Fish Audio releases Drama 3 emotive TTS](https://aifilmhq.com/entry/fish-audio-releases-drama-3-emotive-tts/): Drama 3 is an expressive text-to-speech model in 83 languages that you direct with plain-language cues. It is available in the Fish Audio app and API, and as open weights under a community license. - [HeyGen Video model arrives on Civitai](https://aifilmhq.com/entry/heygen-video-model-arrives-on-civitai/): HeyGen's first general-purpose video model makes 5 to 15 second clips with dialogue, ambience and sound effects generated in the same pass. It is now available in the Civitai generator for text, image or reference to video. - [NVIDIA VFX nodes bring local upscaling, frame generation and HDR to ComfyUI](https://aifilmhq.com/entry/nvidia-vfx-nodes-for-comfyui/): NVIDIA's RTX video effects now run inside ComfyUI as custom nodes on your own RTX GPU. They cover video upscaling up to 4×, frame generation from 2× to 16×, and SDR-to-HDR conversion. - [WanGP v18 adds Kandinsky 6.0 and much faster video VAEs](https://aifilmhq.com/entry/wangp-v18-faster-vaes-and-kandinsky-6/): WanGP v18.00 rewrites the LTX 2, Wan and HunyuanVideo VAEs to be 2 to 3.5 times faster and lighter on VRAM. It also adds the Kandinsky 6.0 audio-video model, Wan 2.2 Prism Video and Wan 2.1 camera control. - [Shengshu launches Vidu Q4 Preview with 4K output](https://aifilmhq.com/entry/shengshu-launches-vidu-q4-preview-with-4k-output/): Vidu Q4 Preview makes video clips up to 16 seconds long at up to 4K with 10-bit color. It accepts up to 15 reference images and 3 reference audio clips to keep characters and voices consistent. - [Kandinsky 6.0 Video released as open audio-video model](https://aifilmhq.com/entry/kandinsky-6-0-video-released-as-open-audio-video-model/): Kandinsky 6.0 Video generates five-second clips with matching sound, including lip-synced speech, from text or an image. The Lite and Pro models, code and weights are released under the MIT license. - [WanGP v17 runs 15 s of MiniMax H3 at 1080p in 11 GB of VRAM](https://aifilmhq.com/entry/wangp-v17-minimax-h3-1080p-in-11-gb-vram/): WanGP, a free local app for open video, image and audio models, rebuilt its memory engine (MMGP v4) in v17. The developer says 15 seconds of MiniMax H3 at 1080p now needs 11 GB of VRAM instead of 25 GB, and 480p fits in 5-6 GB. - [Sync Lipsync v3 available on Artlist](https://aifilmhq.com/entry/sync-lipsync-v3-available-on-artlist/): Sync Lipsync v3 re-renders the mouth in existing footage or a still image to match a new audio track, up to 4K. It holds sync on profile angles and when hands or mics block the mouth. - [Whistle: on-device speech to text in 16.9 MB](https://aifilmhq.com/entry/whistle-tiny-on-device-speech-to-text/): Whistle is an open speech recognition model from Cactus that fits in a single 16.9 MB file and runs on a CPU without dependencies. It transcribes seven languages with word timestamps, which is handy for captions, and audio never leaves the device. - [LTX-2.5 Layout to Render turns a 3D blockout into a finished shot](https://aifilmhq.com/entry/ltx-2-5-layout-to-render-ic-lora/): Lightricks released an open IC-LoRA for its LTX-2.5 22B video model that turns a grey clay viewport or blocky playblast from Blender or Unreal into a finished shot. The layout sets the camera, motion and object placement, and an art-directed first frame sets the look. - [Microsoft ships MAI-Voice-2.1 and Flash](https://aifilmhq.com/entry/microsoft-ships-mai-voice-2-1-and-flash/): MAI-Voice-2.1 is a text-to-speech model covering 23 languages where a single voice keeps its identity across languages. The Flash variant targets high-volume use with about 150 ms end-to-end latency. - [ElevenLabs launches Eleven v4](https://aifilmhq.com/entry/elevenlabs-launches-eleven-v4/): Eleven v4 is ElevenLabs' most expressive text-to-speech model, built for audiobooks, voiceovers, dubbing and character dialogue. A Turbo variant offers about 100 ms latency. - [Use case: a family sci-fi short made with MiniMax H3 on a home PC](https://aifilmhq.com/entry/use-case-short-film-made-with-minimax-h3-on-a-home-pc/): Chris Beach made a short space film starring his family, with every shot generated locally by MiniMax H3 in ComfyUI on one RTX 5090. He published the full workflow, render times and electricity cost. - [Google releases Gemini 3.8 Flash TTS with voice design](https://aifilmhq.com/entry/gemini-3-8-flash-tts-voice-design/): Gemini 3.8 Flash TTS and Flash-Lite TTS let creators design new voices from a text prompt and direct dialogue line by line in over 100 languages. They are available in Google AI Studio, the Gemini API and Google Vids, and every clip is watermarked with SynthID. - [Google Vids offers free HD AI video with Gemini Omni](https://aifilmhq.com/entry/google-vids-offers-free-hd-ai-video-with-gemini-omni/): Anyone with a Google account can now generate 1080p video clips in Google Vids using the Gemini Omni 1.1 Flash model at no cost. Every generated clip carries a digital watermark. - [ComfyUI v0.36.0 adds Video Concatenate and native FastH3](https://aifilmhq.com/entry/comfyui-v0-36-video-concatenate-and-fasth3/): ComfyUI v0.36.0 adds a Video Concatenate node that joins up to 100 clips in a graph, and brings the 8-step FastH3 version of MiniMax H3 into the core build. It also adds Marigold V2 depth, normal and albedo workflows and YuE2 music generation. - [Creatify launches Boreal ad video model](https://aifilmhq.com/entry/creatify-launches-boreal-ad-video-model/): Boreal is a text-to-video and image-to-video model built for advertising that costs one cent per second. It renders roughly in real time and is available via Creatify's API, its apps and fal. - [Nvidia expands AI for Media with lip sync and 8K upscaling](https://aifilmhq.com/entry/nvidia-expands-ai-for-media-with-lip-sync-and-8k-upscaling/): Nvidia's AI for Media toolkit for broadcasters adds frame generation for slow motion, 8K upscaling and SDR-to-HDR conversion. It also adds real-time lip sync to translated audio and speaker detection. - [Nari Labs serves fast, low-cost Qwen3-TTS](https://aifilmhq.com/entry/nari-labs-qwen3-tts-low-latency/): Nari Labs offers optimized versions of the open Qwen3-TTS and Qwen3-ASR speech models through a simple API. The company says they lead Coval's voice AI benchmark on the balance of quality, latency and cost. - [Use case: 'Tusk Hunters' AI short film made with Flick](https://aifilmhq.com/entry/use-case-tusk-hunters-ai-short-film-made-with-flick/): Filmmaker Alan Xiao released Tusk Hunters, an AI short film made with Flick for the Summer '26 Flick Residency Film Series. It features original music by Daniel Lindholm. - [Beeble SwitchX 2.0 video-to-video update](https://aifilmhq.com/entry/beeble-switchx-2-0-video-to-video-update/): SwitchX 2.0 lets filmmakers and VFX artists swap elements in real footage, now with native 4K, 20-second shots and 10-bit color. It also improves camera tracking and lip sync and adds an up-res finishing workflow. - [OpenAI introduces ChatGPT Images 2.5](https://aifilmhq.com/entry/openai-introduces-chatgpt-images-2-5/): ChatGPT Images 2.5 brings sharper detail, more precise editing and faster image generation in ChatGPT. Developers get two API models, GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst. - [Premiere gets a Generative Media Tool in the timeline](https://aifilmhq.com/entry/premiere-gets-a-generative-media-tool-in-the-timeline/): Adobe Premiere can now generate video and sound effects directly in the timeline, using Firefly or partner models such as Veo, Kling, Runway and Luma. After Effects also gets an AI Assistant in public beta. - [Runway plugins for Premiere Pro and After Effects](https://aifilmhq.com/entry/runway-plugins-for-premiere-pro-and-after-effects/): A free Runway panel inside Premiere Pro and After Effects lets editors generate, restyle and upscale footage without leaving the timeline. Generations use the user's existing Runway plan credits. - [Microsoft releases MAI-Image-2.6 and Flash](https://aifilmhq.com/entry/microsoft-releases-mai-image-2-6-and-flash/): MAI-Image-2.6 is Microsoft's strongest image model so far, now available to developers in Microsoft Foundry. A Flash variant targets fast, high-throughput production work. - [LipDub 2.0 speeds up short-form localization](https://aifilmhq.com/entry/lipdub-2-0-speeds-up-short-form-localization/): LipDub 2.0 is a new lip sync model with no training step, making it 8x faster than LipDub 1.0. It handles single-speaker videos up to 2 minutes at 1080p in 80+ languages. - [Use case: 'Xenogenesis' AI short film made with Flick](https://aifilmhq.com/entry/use-case-xenogenesis-ai-short-film-made-with-flick/): Xenogenesis is an AI short film by filmmaker @Gertywood5 made with Flick. It is part of the Summer '26 Flick Residency Film Series. - [fal releases H3 Max video model](https://aifilmhq.com/entry/fal-releases-h3-max-video-model/): fal post-trained the open MiniMax H3 weights into H3 Max for text-to-video and image-to-video. fal says it runs faster than real time and costs $0.08 per second at 768p after the launch promotion. - [OpenShot 4.0 adds color grading and ML subject isolation](https://aifilmhq.com/entry/openshot-4-0-open-source-editor/): OpenShot 4.0 is a major update to the free open-source video editor, adding color wheels, curves, LUTs and video scopes plus built-in screen and webcam recording. It can also isolate subjects using machine learning models that run locally. - [h3-metal: MiniMax H3 running natively on Apple Silicon](https://aifilmhq.com/entry/h3-metal-minimax-h3-on-apple-silicon/): h3-metal (antirez/h3.c) is an open-source C and Metal inference engine that runs MiniMax H3 on Mac computers. It supports prompt-to-video with audio, first and last frames, and image, video and audio references. - [MiniMax H3 open weights run locally in ComfyUI](https://aifilmhq.com/entry/minimax-h3-open-weights-run-locally-in-comfyui/): MiniMax H3 is MiniMax's first open-weights video model, making clips with stereo sound up to 2K and 15 seconds. ComfyUI supported it on day one with pruned, quantized files that can run on a 12 GB RTX 3060.