{"id":20283,"date":"2026-10-01T23:15:04","date_gmt":"2026-10-02T03:15:04","guid":{"rendered":"https:\/\/wp.glbgpt.com\/?p=20283"},"modified":"2026-10-01T23:15:05","modified_gmt":"2026-10-02T03:15:05","slug":"how-to-edit-videos-with-ai-prompts","status":"publish","type":"post","link":"https:\/\/wp.glbgpt.com\/es\/hub\/how-to-edit-videos-with-ai-prompts","title":{"rendered":"C\u00f3mo editar v\u00eddeos con indicaciones de IA en GlobalGPT"},"content":{"rendered":"

Start With Your Footage. Tell Yukie What To Change. Check The Edit On The Timeline.<\/strong><\/p>\n\n\n\n

You have a clip worth keeping, but it needs another shot, a tighter ending, or music that fits. Learning every editing control can turn a small revision into a long afternoon. An AI assistant inside the editor gives you another way to work: describe the change while keeping the footage, preview, and timeline in view.<\/p>\n\n\n\n

To edit videos with AI prompts in GlobalGPT, open Video Editor, import a local file or choose an asset you generated in GlobalGPT, and give Yukie a specific request in the sidebar. Name the clip, describe the change, and say what should stay unchanged. Review the timeline and playback before exporting.<\/strong><\/p>\n\n\n\n

This tutorial uses a bakery comedy assembled in GlobalGPT: an opening shot, a continuation generated from its last frame, and a music track. You will see the full editor, practical prompts to adapt, and the checks that tell you whether an edit actually worked.<\/p>\n\n\n\n

\n
Open Video Editor \u2192<\/a><\/div>\n<\/div>\n\n\n\n
\n

En esta gu\u00eda<\/h2>\n\n\n\n
    \n
  1. What Can You Edit With AI Prompts?<\/a><\/li>\n\n\n\n
  2. How To Start In GlobalGPT<\/a><\/li>\n\n\n\n
  3. How To Write A Useful Editing Prompt<\/a><\/li>\n\n\n\n
  4. Extend A Clip From Its Last Frame<\/a><\/li>\n\n\n\n
  5. Generate And Add Background Music<\/a><\/li>\n\n\n\n
  6. Review And Export Your Video<\/a><\/li>\n\n\n\n
  7. Common Problems And Fixes<\/a><\/li>\n\n\n\n
  8. Preguntas frecuentes<\/a><\/li>\n<\/ol>\n<\/div>\n\n\n\n

    What Can You Edit With AI Prompts?<\/h2>\n\n\n\n

    AI video editing covers several different jobs. A timeline edit changes how existing material plays: cutting, arranging, or adding audio. A generative task creates new material, such as the next shot or a soundtrack. The first can preserve your original frames; the second needs an output review because the model is creating something new.<\/p>\n\n\n\n

    Your Goal<\/th>A Useful Request<\/th>What To Inspect<\/th><\/tr><\/thead>
    Assemble existing clips<\/td>\u201cPut Shot B directly after Shot A, with no gap.\u201d<\/td>Clip order, join, and total duration.<\/td><\/tr>
    Tighten the ending<\/td>\u201cTrim the sequence to 16 seconds. Preserve the opening.\u201d<\/td>The final action is still visible.<\/td><\/tr>
    Continue a scene<\/td>\u201cUse Shot A\u2019s last frame as Shot B\u2019s first-frame input.\u201d<\/td>Actual reference input, framing, and motion across the cut.<\/td><\/tr>
    Add a soundtrack<\/td>\u201cGenerate instrumental comedy music and place it under both shots.\u201d<\/td>A separate audio track and audible exported playback.<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

    These requests cover the operations used in our bakery project. For the wider set of manual controls, including text, transitions, and effects, see the illustrated GlobalGPT Video Editor feature guide<\/a>.<\/p>\n\n\n\n

    You may also encounter \u201ctext-based video editing,\u201d where deleting words from a transcript changes the recording. Descript\u2019s Underlord introduction<\/a> describes that approach. Our example here centers on visual clips, a timeline, and requests to Yukie.<\/p>\n\n\n\n

    How To Start In GlobalGPT<\/h2>\n\n\n\n

    1. Log In And Open Video Editor<\/h3>\n\n\n\n

    Utilice el Video Editor login link<\/a>, or enter from the GlobalGPT homepage. The screenshot marks the two homepage entrances so you can match the route to what you see.<\/p>\n\n\n\n

    \"GlobalGPT<\/a>
    Two Homepage Entrances To Video Editor. Select The Image To View It At Full Size.<\/figcaption><\/figure>\n\n\n\n

    2. Bring In Your Source Material<\/h3>\n\n\n\n

    Import a video from your computer, or select a video you already generated in GlobalGPT. Add the clip to the timeline and preview it before asking for changes. Clear names such as \u201cBakery Opening\u201d and \u201cBakery Continuation\u201d help you refer to the right asset when several versions exist.<\/p>\n\n\n\n

    No source footage yet? Plan a short sequence first, then generate its opening shot. The AI video from script tutorial<\/a> explains how to turn a story into individual visual beats.<\/p>\n\n\n\n

    3. Give Yukie One Clear Editing Job<\/h3>\n\n\n\n

    Open the Yukie sidebar beside the editor. Start with a small request that you can check immediately, such as joining two approved clips. Once that looks right, move to music or a generated continuation. This makes it easier to identify which step changed the result.<\/p>\n\n\n\n

    How To Write AI Video Editing Prompts<\/h2>\n\n\n\n

    A useful editing prompt names five things: the asset, the change, the details to preserve, the timing, and where the result should appear. \u201cMake this better\u201d leaves all five open. \u201cUse Bakery Opening, add the approved continuation after it, and keep the first shot unchanged\u201d gives Yukie an identifiable task.<\/p>\n\n\n\n

    \n

    Reusable Editing Brief<\/strong><\/p>\n\n\n\n

    Use [clip name] on the timeline. [Describe one change]. Keep [the parts that must stay unchanged]. The finished sequence should be [duration and aspect ratio]. Apply the result to the timeline and tell me which clips or tracks changed.<\/pre>\n<\/div>\n\n\n\n
    For a first assembly, use a request like this:<\/p>\n\n\n\n
    \n
    Arrange Two Existing Clips<\/strong><\/p>\n\n\n\n
    Place Bakery Opening first and Bakery Continuation immediately after it on the main video track. Use a straight cut with no gap or overlap. Keep the original playback speed and framing. Trim the end of the second clip so the whole sequence is 16 seconds. Do not regenerate either clip.<\/pre>\n<\/div>\n\n\n\n
    The templates in this guide are edited, reusable versions of the workflow, rather than verbatim chat logs. Replace the example names with your own. A request that works for a simple two-shot sequence may need another pass when your timeline has overlays, narration, or several alternate takes.<\/p>\n\n\n\n
    When the task creates new media, the request also needs a generation scope: which model, how long, and how many attempts. The Gu\u00eda de cr\u00e9ditos de v\u00eddeos generados por IA<\/a> explains why the settings and number of attempts matter to the production budget.<\/p>\n\n\n\n
    How To Extend An AI Video From Its Last Frame<\/h2>\n\n\n\n
    Our example starts in a bakery, with a serious baker and a croissant moving across the counter. The goal was to continue the small visual joke while keeping the same setting and framing. In the September 23, 2026 test, Yukie generated two short shots with Seedance 2.0 and assembled them into a roughly 16-second sequence.<\/p>\n\n\n\n
    The critical step was passing the opening clip\u2019s extracted final frame into the next generation as its first-frame input. Writing \u201cstart where the last clip ended\u201d is less specific: the model also needs access to the actual image.<\/p>\n\n\n\n
      \n
    1. Approve the opening.<\/strong> Check that its final frame shows a usable subject pose and clear scene.<\/li>\n\n\n\n
    2. Request the extracted frame.<\/strong> Tell Yukie to use the actual last frame of that clip as the next shot\u2019s input.<\/li>\n\n\n\n
    3. Describe the next action.<\/strong> Give the continuation one visible story beat, while preserving the setting and camera.<\/li>\n\n\n\n
    4. Review before joining.<\/strong> Compare the new opening with the old ending, then assemble the approved clips.<\/li>\n<\/ol>\n\n\n\n
      \n
      Continue The Bakery Scene<\/strong><\/p>\n\n\n\n
      Use Bakery Opening as the source. Extract its actual last frame and pass that image as the explicit first-frame input for an 8-second continuation. Keep the same baker, croissant, counter, lighting, camera position, and 16:9 framing. Continue the action: the croissant slides toward the waiting plate and the baker notices it. No camera cut, text, dialogue, or new characters. Use a model that supports first-frame input. Generate one continuation, then place it after the opening for review.<\/pre>\n<\/div>\n\n\n\n
      \"Full<\/a>
      The Bakery Project: Assets, Preview, Both Video Clips, And Yukie In One Full Editor View. Select To Enlarge.<\/figcaption><\/figure>\n\n\n\n
      The Combined Bakery Sequence. This Silent Version Was Derived From The Editor Export To Make The Visual Join Easier To Review.<\/figcaption><\/figure>\n\n\n\n
      What worked:<\/strong> Yukie used the extracted frame as an explicit input, generated the continuation, and assembled the two shots. The scene remained recognizably the same bakery. What still needed judgment:<\/strong> the intended plate landing and eyebrow reaction did not fully read in the output. Supplying a frame does not guarantee a pixel-identical join or exact performance.<\/p>\n\n\n\n
      Review the cut at normal speed as well as frame by frame. A close visual match can still feel awkward if the croissant changes speed or the camera jumps. For more continuity checks, use the guide to keeping AI video characters, props, and scenes consistent<\/a>.<\/p>\n\n\n\n
      How To Generate And Add Background Music To A Video<\/h2>\n\n\n\n
      Once the edit has a stable duration, describe the soundtrack in terms of mood, instruments, and timing. For the bakery scene, the brief called for an instrumental comedy cue with pizzicato strings, bassoon, and soft woodblock. Yukie used Eleven Music V2 in the same test and placed the result on a separate audio track.<\/p>\n\n\n\n
      \n
      Create Music And Add It To The Timeline<\/strong><\/p>\n\n\n\n
      Create a 16-second instrumental comedy cue for the bakery sequence using an available connected music model. Use playful pizzicato strings, a light bassoon melody, and soft woodblock. Keep it understated, with no vocals. Import the generated audio and place it on a separate track starting at 00:00. Mute the original audio in both video clips and trim any music overhang to match the sequence.<\/pre>\n<\/div>\n\n\n\n
      Generating an audio file is only the first half of this job.<\/strong> Confirm that a waveform appears underneath the video and spans the intended part of the edit. In our workspace screenshot, the purple audio clip sits below both shots, and its properties are open.<\/p>\n\n\n\n
      \"Full<\/a>
      Music On Its Own Timeline Track, Beneath The Two Video Clips. Select To Inspect The Full Workspace.<\/figcaption><\/figure>\n\n\n\n
      \n
      Listen To The Generated Comedy Cue<\/strong><\/figcaption><\/figure>\n<\/div>\n\n\n\n
      If your footage has speech, adjust the example: ask to retain the original dialogue and lower the music underneath it. Listen to the exported file to judge the balance; a visible waveform confirms a track exists, but does not prove that the mix is right.<\/p>\n\n\n\n
      For examples of how arrangement and timing instructions affect a generated track, read our ElevenLabs Music V2 review and prompt tests<\/a>.<\/p>\n\n\n\n
      Review The Timeline Before Exporting<\/h2>\n\n\n\n
      Keep the preview, timeline, and Yukie\u2019s response in view together. The response explains the intended work; playback and the tracks show what is actually in the edit. In this project, the real exported file was about 16 seconds, at 1920 \u00d7 1080 and 30 fps, with a non-silent audio track. Those are this test\u2019s export properties, rather than a promise for every model or account.<\/p>\n\n\n\n