Artificial Intelligence

    AI videos from script to final cut.

    AI videos for products, campaigns, websites and training start with your brief. We write the script, generate images and motion, edit with music and sound and deliver every format your channels need.

    For marketing, communications and training teams who need clips for trade fairs, social media, websites or courses, with brand and rights settled.

    • Grok Imagine
    • Higgsfield
    • Kling
    • Claude
    • ffmpeg
    The arcades of the Piazza Grande in Locarno, a man edits a video on his laptop, a clapperboard, a camera on a tripod, a storyboard and the robot agent beside him
    01

    From data sheet to trade-fair clip

    The product manager sends a data sheet, three product photos and one sentence: the clip should run on the trade-fair stand and on LinkedIn. From this we build a script and storyboard, scene by scene with an image idea and text. She cuts one scene and approves the rest. At the end there are three versions: wide for the stand screen, square for the feed and vertical with subtitles for the phone.

    02

    What we produce for you

    Your brand, your products and your messages stay the foundation. We combine them with models that generate images and motion, and with the craft of editing, music and sound. Logo, type and product footage go in during the edit, so they are exact in every frame. Every video goes through your approval before delivery.

    • Product and explainer clipsA new product, a feature or a process, shown in scenes that would be hard to film with a camera, such as a look inside a machine.Result: a clip for website, trade fair and sales, with subtitles in your languages.
    • Social and campaign videosVariants of one idea for LinkedIn, Instagram or YouTube, in the right aspect ratio and with an opening that holds attention while people scroll.Result: a campaign of several cuts your team can test against each other.
    • Hero and background loops for websitesCalm scenes that repeat seamlessly or play once and stay on the last frame, with a poster image and a lean file.Result: a video that loads fast on a phone and stays sharp on large screens.
    • Training videosA process from your daily work, explained step by step, with a voice-over, screen recordings and scenes that show the context.Result: short chapters your staff open in the intranet, in Teams or in the learning system.
    • Avatar and voice-over videosA person from your company speaks as an avatar in several languages, or a voice-over guides through the film. We only work with people who have given written consent.Result: the same message in German, French and English, with the same person.
    03

    What we need from you

    A brief with goal, audience and channel, your brand guidelines with logo, fonts and colours, product photos or screenshots that appear in the video, and the statements you have approved. Anyone who appears in the video or lends their voice gives written consent for purpose, channels and duration. And one person on your side who approves storyboard and edit.

    04

    What you hold at the end

    The finished videos in the agreed formats, with subtitle files and a poster image per clip. Plus the storyboard, the prompts used and a rights sheet: which tool on which plan, which terms of use with their date, which music licence and which consents cover each clip. We label every video as made with AI, in the credits, in the description or in the picture, whatever suits the channel.

    From our own workshop

    How the video above was made.

    It runs on our start page: the group crossing the bridge to the hut is made of generated watercolour clips. We then lifted the slightly yellowish paper to pure white so the video blends into the page, scaled it to 1080p, sharpened it lightly and encoded it for the web.

    The drawn pictures at the top of our expertise pages are made the same way: one scene per page, described in a prompt, generated and then edited by us until white space and format fit the page.

    Sound familiar? Tell us where you stand today.

    Contact

    Usage rights and availability.

    As of: 7 November 2026

    • As of 27 October 2026

      Grok Imagine: output belongs to the customer, with consents and attribution.

      The enterprise and API terms assign all rights in the output to the customer. The terms for personal accounts carry specific rules for Europe and Switzerland and require the necessary consents for every face, voice and image. The personal-account terms also require attributing the output to the service, as xAI's brand guidelines describe.

      Source: xAI: Enterprise Terms of Service
    • As of 23 October 2026

      Higgsfield claims no rights in your videos.

      The terms of use state that Higgsfield owns neither inputs nor outputs; according to the help centre, commercial rights apply on every plan, and downloads without a watermark come with paid plans. Showing people requires their consent. Switzerland is not excluded; processing takes place in the US.

      Source: Higgsfield: Terms of Use
    • As of 10 October 2026

      Kling: commercial use for paying members.

      The terms of paid service do not restrict commercial use of the output by members, and members may remove the watermark. Without a membership, the general terms require written permission and the Kling logo. According to the privacy policy, data is stored on servers in Singapore.

      Source: Kling AI: Terms of Paid Service
    • As of 7 November 2026

      Claude: output rights with the customer, Switzerland supported.

      Anthropic's commercial terms assign the rights in the output to the customer, and customer data is not used for training. Switzerland is on the list of supported countries. Processing takes place in the US or globally; the first-party API offers no processing in Europe.

      Source: Anthropic: Commercial Terms of Service
    • Since 2 August 2026

      In the EU, deepfakes must be recognisable as AI content.

      Article 50 of the AI Act obliges anyone deploying deepfakes to disclose that the content is artificially generated or manipulated; for artistic works a note that does not hamper the work is enough. In Switzerland the data protection act applies directly to AI, and the data protection commissioner expects manipulated faces and voices of identifiable people always to be clearly recognisable as such.

      Source: EUR-Lex: AI Act, consolidated text

    How a video is made.

    Every step ends with something you see and approve before the next one starts.

    • Brief

      Goal, audience, channel

      We settle with you what the video should achieve, who sees it and where it runs. Length, aspect ratio, languages and tone follow from that.

    • Script & storyboard

      Script, scenes, prompts

      Claude writes the script and the scene order with us, the image and video prompt for every scene and the editing logic: which shot stands for how long, where text and logo appear. You approve the storyboard before a single image is generated.

    • Images & motion

      Grok Imagine, Higgsfield and Kling

      Each scene is made in the tool that carries it best, such as drawn and painted worlds, camera moves, characters or people speaking. We generate several variants per scene and choose with you.

    • Edit & sound

      Music, voice, subtitles

      In the edit we place logo, type and real product footage, lay music and voice-over underneath and match colour, brightness and sharpness across all scenes.

    • Delivery

      Every format for your channels

      16:9 for website and presentations, 1:1 and 9:16 for social media, as MP4 for the web with a poster image, plus a master file in full quality and the subtitles as a separate file.

    What marketing and communications ask first.

    May we use the videos commercially?

    Yes. We produce on paid plans whose terms allow commercial use of the output, and the rights sheet records for every clip which tool, which plan and which terms it was made under, with the date. We license music for your channels. That way your legal team can trace every clip on its own.

    Can someone from our team appear in the video?

    Yes, with the person's written consent for purpose, channels and duration, and with the right to withdraw it. That applies to face and voice alike. We do not recreate people who have not consented, public figures or characters from other people's works.

    Do you label AI content?

    Yes, every video. Where and how we agree with you per channel: in the credits, in the accompanying text or as a note in the picture. Provenance data a tool writes into the file stays in. For audiences in the EU the AI Act requires deepfakes to be disclosed, and the Swiss data protection commissioner expects the same for altered faces and voices; we label regardless.

    Do logo and product stay exact?

    Yes, because we do not let them be generated. Video models change fine type and logos from frame to frame; we have seen it in our own videos. Logo, type, prices and real product footage therefore go in during the edit, and the generated scenes carry mood, place and motion.

    Which data goes to the vendors?

    Prompts, reference images and the text a scene needs go to the tool in question and are processed there, outside Switzerland, for example in the US or in Singapore. Confidential material such as unreleased products or personal data we only upload with your approval; a neutral reference image is often enough. The rights sheet records which data went to which tool.

    What if we don't like a scene?

    Then we change it, ideally in the storyboard before an image is generated. Generated scenes can be regenerated one by one without touching the rest. How many rounds of changes the order includes is set out in the offer.

    What do we get besides the finished videos?

    Storyboard, prompts, the selected scenes as separate files, the master file, subtitles and the rights sheet. With these your team can later create a new language version or a shorter cut itself, or order it from us.

    Agree on a useful next step

    Tell us about your current workflow, the systems involved and the result you need. In our first conversation, we identify what else we need to know to plan a useful next step.