> For the complete documentation index, see [llms.txt](https://thecontentforge.gitbook.io/thecontentforge-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://thecontentforge.gitbook.io/thecontentforge-docs/feature-glossary/glossary/video-forge.md). # Video Forge *Last updated: 2026-08-07.* Video Forge turns one long recording into short-form video and the written content that goes around it. You upload an interview, podcast, webinar, or property tour once, and the workspace transcribes it, finds the moments worth cutting, and gives you a full editor that produces finished, captioned, platform-shaped clips. * **Video upload** - Drag a video in or browse for one, in the common recording formats and at any length. Progress is shown live and can be cancelled mid-upload, and the panel tracks how many videos your plan allows this month so you always know what is left. * **Recent sessions** - Every video you have processed stays on the landing screen with a thumbnail, its status, its length, and the date it was added, so returning to an old recording never means uploading it again. * **Processing progress** - While a video is being prepared, a five step card shows exactly where it has got to, from upload through preparing the video, captions and analysis to ready to edit. If a stage does not complete, that stage is marked and the reason is shown rather than leaving you with a stalled spinner. * **History switcher** - Inside a session, one control lists every other session with its status, length, and how many posts it has already produced, and jumps to it or starts a fresh upload without going back to the landing screen. * **Automatic transcription** - When processing finishes, a timestamped transcript is produced automatically and sits next to the player. You can search it, copy it in full, and click any line to jump the video to that exact moment. The transcript is also what every insight, clip suggestion, and generated post is grounded in. * **Transcript retention** - The panel states how long the transcript stays available after upload, so you know when to export anything you want to keep. * **Word-level caption timing** - Every individual word in the range you are working on is timed, which is what lets captions highlight word by word in sync with speech, and timing is ready quickly for the clip you are cutting. * **Insights tab** - Reads the whole transcript and returns a plain summary, the key topics, the strongest quotes with speaker and timestamp, the best hooks, the key moments, and follow-up angles worth posting later. Quote and moment timestamps are clickable, so you can hear a quote in context before you commit to it. * **Visual tab** - Covers what is on screen rather than what is said: on-screen text, charts and graphics with captions, chapter breaks, a frame-by-frame description, accessibility alt text, and its own clip suggestions. This is what catches the value in screen shares, slide decks, and demos where the words alone tell half the story. It can be re-run at any time, and it is available for as long as the source video is still kept. * **Clip suggestions and custom windows** - Suggested clip windows can be cut as they are, nudged to a different start and end, or replaced entirely by a window you type in yourself with your own label. Start and end times have to sit inside the actual length of the video, and each cut becomes a real, downloadable video file rather than just a timestamp. * **Content tab** - The at-a-glance view of a finished session: what the video is, who is speaking, why it is worth sharing, the summary, the top three quotes, every clip you have cut, and every post generated from it, all in one place. * **Clips list actions** - Each cut clip carries its own row of actions: generate copy from it, send it straight to a post, download it, or save it to your asset library. A clip whose cut failed offers a retry on the same row. * **Generate tab** - Pick from fifteen output types, including a single post, a thread, a quote card, key takeaways, platform captions, teaser copy, clip titles and hooks, an episode title, show-notes style description copy, and thumbnail text. Choose Fast for speed or Best for quality, select as many outputs as you want, and everything is written against your brand voice. Only the output types that make sense for the length of the video are offered, so a thirty second clip never gets an episode summary. * **Generation status** - While outputs are being written, each one is listed with its own live state, so a single failure is visible on its own line instead of sinking the whole batch. * **Generate from a clip** - Any clip can be sent to the generator so the copy is grounded strictly in that window rather than the whole recording, and the clip video attaches automatically when you publish. The pre-ticked outputs then follow the clip's length, not the full recording's. * **Vertical presets** - Student accounts get one-click output bundles tuned to what they are doing, such as a game day pack, an event recap pack, or an episode promo pack, so they do not have to work out which outputs belong together. * **Create tab** - The clip production workspace: AI clip suggestions, the clips you have started, the full editor, the render queue, and your library of finished renders. It opens once the session has a playable source video. * **AI clip finder** - Ask for the best moments automatically, or describe what you are looking for in your own words and get only the moments that match. You can set a target clip length, output shape, platform preset, safe area guide, spoken language, and whether a hook is added on top. * **Clip scoring** - Every suggested clip carries a hook score, a virality score, and a brand fit score alongside the transcript excerpt it came from. It gives you a defensible reason to cut one moment over another instead of guessing. * **Clip preview** - Before committing, play the suggested window on its own, read the score breakdown and a short explanation of why it works as a standalone clip, then either create the clip or open it straight in the editor. * **Edit the whole video** - You can also open the full source recording in the editor without picking a suggestion, which is the route for captioning or trimming a short video end to end. * **Clip editor** - A full-screen editor with a live preview that matches the final export, a multi-lane timeline, layer controls, undo and redo, keyboard shortcuts for play, split, duplicate and delete, and continuous autosave. What you see in the preview is what renders, so there are no surprises at export. * **Caption styles** - A browsable grid of caption looks with live previews, grouped by category, covering font, weight, size, casing, text and highlight colours, the highlight pill behind the spoken word, outline stroke, keyword colouring, emoji on keywords, and line backgrounds. A style looks identical in the preview and in the finished file. * **Caption line control** - Set how many words appear per line and how many lines show at once, which is the difference between captions that read cleanly on a phone and captions that cover the speaker's face. * **Caption positioning** - Place captions at the top, centre, or bottom, nudge them away from or toward the edge, set an exact custom height, or simply drag the caption block on the preview to where you want it. * **Saved caption styles** - Once you have a look you like, save it by name and apply it to any future clip in one click, which keeps a whole library of clips visually consistent. * **Caption consistency on older clips** - New clips get the current caption behaviour, with lines broken across the screen properly and captions sitting exactly where the preview shows them. Clips you started earlier keep the caption look their existing exports already have, and the editor tells you so; editing their captions brings them onto the current behaviour. * **Transcript editing and text-based cutting** - Edit the video by editing its words. Click a word to jump there, fix what a caption displays, hide words from captions without touching the audio, or select a range and turn it into a clip, cut it out, or split there. Long pauses and filler words can be detected and removed in a reviewable pass, and nothing you do here alters the original recording. * **Find similar moments** - Select a line in the transcript and ask for more moments like it, and the clip finder comes back with matching windows from elsewhere in the recording. * **Auto-reframe** - Turn a wide recording into a vertical clip that follows the speaker instead of statically cropping the middle of the frame. Manual pan and zoom controls are there when you would rather frame it yourself, and manual framing overrides the automatic track. * **Multi-panel layouts** - Beyond a single filled frame, a clip can be laid out as a split, a three or four way grid, a screen share, or a gameplay arrangement. Each panel is framed independently, so you choose what shows in the top half and what shows in the bottom. * **Source as a second layer** - The source clip can be duplicated into a free-floating, croppable layer over its own window, which is how you put a cropped facecam on top of full-frame gameplay or a screen share. * **Cuts and transitions** - Split a clip at the playhead, extend a cut out to the nearest sentence boundary, and choose what happens at each join rather than accepting a hard cut everywhere. * **Stock footage** - Search a large stock library for video and photos from inside the editor, browse the thumbnail results, and import the ones you want straight onto the timeline as B-roll. Imported footage is ready to play on the timeline and is credited to the original creator. * **Media layers** - Add your logo, images, banners, and your own B-roll video, either uploaded fresh or pulled from your existing asset library. Each item is its own layer that you drag, resize, and crop directly on the preview, with fit and fill options, per-item colour adjustment, and entrance and exit animation. * **Layer list** - Every visual layer, plus the source video and the captions themselves, appears in one ordered list you can drag to change what sits in front of what. Selecting a row selects that layer on the preview. * **Text, hooks, and end cards** - Add hook text at the top of a clip, titles, callouts, and closing cards that automatically land at the end of the clip. A one-click AI hook drops in a scroll-stopping opening line drawn from the video itself. * **Forge Moves** - Motion for any visual layer or the source footage, either from a set of ready-made movement presets or from manual keyframes with easing when you want exact control. One click can also add a punch-in or a smooth push at every cut point to soften jump cuts. * **Look and templates** - Switch the output shape, choose a background treatment behind the video, and turn on finishing effects such as film grain, vignette, cinematic bars, a progress bar, light leaks, and gradient backdrops. Style packs and templates apply a coordinated look across captions, transitions, and text in one click. * **Audio building** - Add music, sound effects, and voiceover from an upload or from your asset library, control volume, fades, and trims per track, and duck the music automatically under speech so the dip follows the voice naturally rather than dropping to a flat level. Beat detection marks the music's beats on the timeline so cuts can land on them. * **Enhance speech** - A light or strong cleanup of the spoken audio that rescues clips recorded in noisy rooms or on weak microphones. It applies to the exported file rather than to the editor preview, so you hear it in the finished clip. * **Loudness normalization** - An optional export setting that brings the finished clip to the standard loudness level social platforms expect, so your clips do not play noticeably quieter or louder than everything else in the feed. * **Rendering and export** - Render a finished clip and watch its progress in a queue. You can choose final or draft quality, burn captions in or leave them off, match the source frame rate automatically or pin it, and add a watermark with your own text. When it is done you can play it in the browser or download the finished video file. * **Aspect variants** - Duplicate a finished clip into another shape, vertical, wide, square, or portrait, so one edit can serve several platforms without rebuilding it. * **One-click auto-caption** - From the clips list, pick an output shape and get a fully captioned video back without ever opening the editor. When the shape you pick differs from the source, it also reframes to follow the speaker rather than cropping the centre. It carries on if you leave the tab, and you can still open the result in the editor afterwards to adjust the styling. * **Rendered clips library** - Every finished render is kept with a thumbnail, its shape and length, and the date it stays available until, with play, download, and delete on each one. Anything past its window can be re-rendered. * **Save to library** - Push a clip into your workspace's asset library so it can be attached to scheduled posts like any other piece of media, and so it stays available after the original source video has passed its retention window. * **Publish and revise generated posts** - Every generated post can be copied, sent back for a revision with your own instruction, or published and scheduled directly from the card, with the matching clip attached. * **Thumbnail remix** - Take a frame from the video and restyle it into a thumbnail by describing the change you want, then keep iterating on the result until it fits the post. * **Cancel or delete a session** - Stop work that is still running or remove a finished session outright, including its transcript, insights, and drafts. A session already attached to a pending post is protected until that post is dealt with. * **Property Video Forge** - Real estate accounts get the same workspace framed around property tours, agent updates, testimonials, and market briefings, with a direct route into building a video from a property workspace. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://thecontentforge.gitbook.io/thecontentforge-docs/feature-glossary/glossary/video-forge.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.