From first idea
to final cut.

CineGen brings planning, generation, editing, and review into a project. You can start with footage you already have or create new footage with AI.

1. Sign in, open Dashboard, and choose New Project. Enter a project name, choose the aspect ratio and resolution offered for your plan, and create it.
2. Open the project’s Editor. Add a video, image, or audio file to Media, then drag it onto a timeline track.
3. Select the clip. Drag its edges to trim it, move it to set its start time, and press Space to preview the cut.
4. Add music or a text asset on another track. Select each item to adjust it in the Inspector.
5. Choose Export, set a filename, and export a draft. Open the downloaded file to check the picture and sound.

For an AI-first project, create your cast in Character Creator, plan shots in Storyboard, then use Scene Creator → 2D scenes to generate and continue clips. Each generated clip becomes media you can arrange in the Editor.

Use the workspaces together or enter at the stage your project needs. A storyboard shot is a plan; a generated clip is a media asset; a timeline clip is an instance of that asset in your edit.

Help opens this guide. Support opens your support conversations. A locked action usually needs a different plan, sufficient tokens, or permission to edit that project.

A project contains the production. Scenes separate its shots, timelines, and creator work, so a new scene does not need a new project.

1. Create or open a project on the Dashboard. Use the pencil beside its name to rename it.
2. Use the scene picker in the project to add or select a scene. Check the current scene before generating, editing, or sharing.
3. In Storyboard, give the scene a Sequence / episode, Location, and Frame. Use Project scenes to jump to a scene’s storyboard, timeline, or creator.
4. Use explicit save controls where shown: Save shot in a storyboard panel and Save in the 3D workspace. Wait for saving or uploading to finish before closing the tab.
5. Reopen the same project and scene to continue working. If a collaborator has saved a newer version, review the conflict message before replacing your draft.

Free editing stores timeline and imported media locally in this browser. Cloud project details do not mean those local files are backed up. Keep source files and exported copies; changing devices, private browsing, or clearing site data can make local work unavailable.

You can read this guide without an account. In the app, your subscription, token balance, and project role determine which actions are available.

A workspace preview does not grant permission to run every action in it. Owning tokens alone does not unlock all AI tools. The app displays upgrade prompts and costs at the action you are about to run.

Open Pricing for current prices and allowances, then use Settings → Billing to manage your own account.

Build a saved identity before generating a series of shots. A clear portrait and specific physical description give later tools a stronger reference.

1. Open Character Creator. Upload a portrait or open AI Image Generator and describe the character you want.
2. Enter Character Name, choose a category, and fill in Personality Traits and Clothing & Style.
3. Use Analyze Image to develop the Physical Description (from AI), then correct anything inaccurate. You can also use the AI profile generator to develop character details.
4. Choose a voice if the character will speak, then select Save Character.
5. Find the character in the searchable Character library. Use categories to organize a growing cast and open a saved character to create references.

Describe identity with stable details: face shape, hair, build, distinguishing features, and outfit. Save scene lighting and camera movement for the shot prompt. Generated reference lighting is kept neutral so the same character can be used in different scenes.

A wardrobe sheet ties an outfit to the saved character. It helps distinguish a deliberate costume change from an unwanted change of identity.

1. Open a saved character and describe the outfit in its wardrobe tools. Give the outfit a recognizable name and choose Create sheet after checking the token cost.
2. Inspect the generated sheet for facial identity, clothing, colors, and accessories. The reference brings together front, back, and face views.
3. Select the outfit as Default for new scenes, or choose Original look to use the original portrait.
4. Use Facial expressions and Angle shots to generate additional references. Open a result full-screen to inspect it before relying on it.
5. In a storyboard shot, add the character and record the costume and hair under Character continuity. Match the costume name to the saved wardrobe you intend to use.

Changing a default wardrobe affects new uses of the character. It does not rewrite existing rendered footage. A 2D continuation keeps its source clip’s cast and wardrobe; start a new shot for a deliberate cast or outfit change.

Save a voice on the character to reuse it in dialogue generation and voice changes.

1. Select a voice under Character Voice (ElevenLabs) and play the voice preview.
2. Adjust the character’s Voice Effects if you want a reusable vocal treatment. Listen to the preview before saving.
3. Select Save Character. Use that character when generating dialogue from a storyboard or the Editor.
4. To change an existing clip’s voice, select it in the Editor, open Enhance → Character voice, choose a character with an assigned voice, and run the action after reviewing its cost.
5. Listen to the full result for pronunciation, timing, and sound quality. Mix its volume with music and other dialogue in the Editor.

A selected visual character does not automatically guarantee its assigned voice was used. Check the generation panel’s dialogue and voice options before submitting.

A verified avatar is a separate consent and liveness workflow for making a reusable on-camera version of yourself.

1. Open the verified-avatar setup while signed in to your own account and read the consent presented there.
2. Follow the capture instructions. Allow camera access and complete the requested liveness steps in clear, steady light.
3. Submit the capture and wait for the verification result. Use the capture handoff offered in the app if you need to record on another device.
4. Once verification succeeds, select the resulting avatar in Talking Head. Choose an appropriate wardrobe reference when available.

If capture fails, check camera permission and lighting and follow the displayed retry instructions. Contact Support for a verification or ownership issue. A normal uploaded portrait is not a substitute for verified-avatar approval.

Use a storyboard to decide the action, framing, dialogue, and order before committing to a render.

1. Choose the scene, then select Add shot. Enter a shot name, Shot type, Duration (s), Action / visual direction, Dialogue, and Camera movement.
2. Add a reference still or link project assets. Add the cast under Character continuity and record the intended look.
3. Choose Save shot. On the board, drag a shot onto another shot in the same scene to reorder it, or use the earlier/later arrow buttons.
4. Use Show → Whole project, Group by, and Find shot to scan a larger production by scene, episode, or location.
5. Use Timeline or Creator on a shot card to open its scene in the next workspace. Archiving a shot preserves existing timeline and creator work; Archived shots → Restore shot brings the plan back.
6. For an earlier revision, open the shot’s Version history, choose Load as draft, inspect it, then Save shot to restore it. If a collaborator saved a newer version, review the conflict and use Reload latest shot when you want their version.

Planned duration is an editorial target. It can be longer or shorter than the fixed durations offered by a video-generation model.

Start from a screenplay or paste a draft directly into CineGen. Script import and manual breakdown do not use AI credits.

1. Choose Import script for Final Draft (.fdx), PDF, Word (.docx), Fountain, or plain text. Alternatively, paste into Script.
2. Select Save script & refresh breakdown after editing the text. Review the detected scene headings, action, speakers, and dialogue.
3. Turn on Create project scenes from headings if the breakdown should create separate project scenes.
4. Select Add breakdown to storyboard to add the suggested shots. To be more selective, use Tag shot beside an element or highlight text and choose Tag selected text as shot.
5. Open the resulting shots to refine framing, timing, references, and continuity. Script links retain the relationship to their source text.

Check the imported text before adding the breakdown, especially for unusual PDF layouts. If text is missing or scrambled, paste readable script text and refresh the breakdown.

Use the panel canvas to communicate composition without needing a finished image.

1. Choose a Reference still or Upload reference if you want to draw over an image. Adjust Reference brightness or Grayscale to make your marks legible.
2. Select a Tool: vector pen, bitmap brush, bitmap eraser, line, arrow, rectangle, ellipse, or text/callout. Set ink and width before drawing.
3. Choose a thirds or perspective Guide for composition. Use Move mark and Selected vector mark to reposition vector marks, delete them, or bring them to the front.
4. Choose Camera path, draw the move, and set its zoom. Select Preview camera move to inspect it; Clear camera path removes it.
5. Use Undo or Redo while editing, then Save shot. Export an individual panel as PNG · 1920px or TIFF · 3840px.

The bitmap eraser affects bitmap marks. Vector marks have their own selection and deletion controls. Saved drawings and references can become the visual guide for a 2D generated shot.

Continuity groups describe which shots should share a look. Use a new group when the story intentionally changes costume or hair.

1. Open a shot and use Add character to choose someone from the character library.
2. Enter a Continuity group, Costume, Hair / makeup, and any Continuity notes.
3. Choose Match saved costume / hair to reuse the look already recorded for that character and group.
4. Review Continuity differences before saving. The Continuity tab summarizes the cast across shots and scenes so you can find mismatches.
5. Carry the shot into Scene Creator and review the actual references before generating.

These notes and images guide generation; they do not guarantee an identical result. Inspect faces, hands, accessories, screen position, and lighting at every join before continuing a sequence.

AI can draft your shot plan, and individual shots can generate media after you refine that plan.

1. Open AI generation, enter Story direction, choose characters, and select any project style prompts.
2. Review the displayed cost and select Generate sequence. The generated shots are added to the current scene. Inspect and refine their action, order, and dialogue before rendering footage; use Previously generated sequences to add a saved sequence again when needed.
3. Open a saved shot and choose Generate video · credits to render it, or Generate voice · credits for its dialogue. Supply an assigned character voice where required.
4. Find the output in the project’s media and bring it into the Editor. For iterative shot development with cast and storyboard-image controls, choose Open shot in scene creator instead.
5. For multiple shots, open Project scenes → Batch Video Generation, select the shots and offered durations, check the total estimate, and start the batch. Check each shot’s result independently.

AI planning and media generation are separate paid actions. Review and correct a text plan before using tokens to render its footage.

Choose the board scope first: exports use the shots currently shown, so scene, grouping, and search choices matter.

1. Open Export & timeline beneath the board.
2. Choose Shot list CSV for a spreadsheet or Structured JSON for structured planning data.
3. For a printable board, set PDF title / company, Watermark, Panels / page, and PDF frame, then select Export PDF.
4. Choose MP4 animatic to preview the planned shot durations and drawn camera paths as a video. Keep the tab open during the local export; use Cancel if needed.
5. With Show → Current scene, choose Add shots to this timeline to append the storyboard as a new timeline track. Existing edits remain in place.

An animatic visualizes the plan. It is not a newly generated live-action performance. Use it to judge pacing and composition before rendering final footage.

Use the production plan to connect scenes with shooting days and crew information.

1. Open Production and add a shooting schedule row.
2. Set the Date, Scene, Call time, estimated Minutes, Location, Cast, and Notes.
3. Add Crew contacts with a name, role, contact information, and call time.
4. Select Save production plan. Use Export schedule CSV for the full schedule or a Call sheet button for the shooting day you need.
5. Check the exported dates, scene names, addresses, and contact details before distributing it.

The schedule is a production document. Entering a crew contact does not invite that person to your CineGen project.

The 2D workspace builds a scene one clip at a time on a single track. “2D” describes this generation workspace; you can direct live-action footage, animation, or another visual style.

1. Select your project scene and choose New shot. Describe the visible action under What happens in this shot?
2. Select saved characters and, if useful, a storyboard shot. Open Staging, camera & continuity to specify location, screen positions, camera, lighting, style, exact dialogue, and the clip name.
3. Choose duration and aspect ratio. Text-only shots offer 4, 6, or 8 seconds; shots using references use 8 seconds. Available frames are 16:9 and 9:16.
4. Choose Review director prompt to inspect the compiled direction, references, and cost. Edit the fields until they describe one clear shot.
5. Generate and wait for the result. The finished clip appears in the scene’s sequence and the project’s media bin. Use Play sequence to watch the clips in order.

Keep generation here and detailed editing in the Editor. The sequence does not provide the Editor’s trimming, transitions, audio mixing, or multi-track composition.

Continuation uses the saved endpoint of a prior generated clip, along with its visual and character context, to guide what happens next.

1. Watch the clip you want to continue and select Continue this clip.
2. Under What happens next?, describe the next physical action. Keep the same cast, wardrobe, location, and screen direction unless the action itself moves them.
3. Review the prompt and cost, then generate the 7-second continuation. It is appended to the single-track sequence and saved to Media.
4. Watch the join, not just the new clip. Check eye lines, object positions, movement, costume, and sound before continuing again.
5. Repeat until you have the footage you need. Use New shot for a different setup or cast. Open the scene’s Editor to assemble the final cut from the media bin.

You can continue a selected earlier clip; the result still appends to the sequence. Use Remove from sequence for an unwanted entry. Removing an entry does not delete its media asset or undo a generation charge.

Use the shot plan as the starting point for a directed generation instead of retyping all its details.

1. Open a shot with Creator or Open shot in scene creator, or select it from the 2D workspace’s storyboard control.
2. Review the imported action, dialogue, camera direction, location, and cast. Correct anything that should differ in the generated shot.
3. Enable the storyboard-image option if its saved drawing and reference should guide composition. It uses the rendered panel, including the visual marks you saved.
4. Select the character references. A new shot supports up to three character references, or two characters plus the storyboard image. Check the intended wardrobe.
5. Open Review director prompt to inspect what will actually be sent. Generate only after the image roles and written direction agree.

A storyboard image guides composition; a character reference guides identity and outfit. Keep the action prompt clear about which person is where. For a continuation, the source clip’s cast and wardrobe take priority.

Describe what the camera will see in the next few seconds. Concrete actions, a simple camera move, and clear geography are easier to follow than a list of moods.

1. Type your direction, or choose Dictate in a supported browser and allow microphone access. Stop dictation and correct the text before generating. If dictation is unavailable, all the same controls work with typing.
2. Use Develop direction to expand the idea into action, staging, camera, and lighting guidance. This is a separate text-generation action with its own displayed cost.
3. Define positions using screen-left, screen-right, foreground, and background. State what each person holds and which way they look or move.
4. Choose one achievable camera instruction. Name the light source, keep style consistent, and put spoken words in the exact-dialogue field.
5. Review the final prompt and references. Remove contradictions, unnecessary actions, or text you do not want visible on screen.

Example: an opening shot

Mara stands screen-left beside a stone railing, holding a brass telescope in her right hand. The lighthouse stays in the background on screen-right. She raises the telescope to her eye. Medium shot, slow push-in, warm light from the setting sun at camera-left. Hand-painted animation, muted sea-green palette.

Example: the continuation

Continue from Mara’s final pose. She lowers the telescope to chest height, keeping it in her right hand, then turns her eyes toward the lighthouse. Keep her on screen-left, the railing in the foreground, and the lighthouse on screen-right. Maintain the same camera axis, costume, animation style, and sunset lighting.

Project style prompts can carry a recurring visual treatment into generation tools that offer them. They supplement the shot direction; review them for conflicts with the reference imagery.

Project styles store written direction you want to reuse, such as a color palette, animation medium, lighting approach, or lens treatment.

1. Open Project styles in a generation panel that offers it.
2. Under New style, enter a short name and describe the lighting, color, texture, lens, medium, and mood. Choose Add style.
3. Select one or more saved styles for the current generation. Creating a style and selecting it are separate actions; check the selected count.
4. Use the pencil control to change a saved style and choose Update style. Deselect a style when it does not belong in this shot.
5. Review the combined direction before generating. Conflicting styles can weaken the result, particularly when they disagree with a character or storyboard reference.

Styles are appended to the generation where you select them. Editing a saved style does not change footage you have already rendered.

Use 3D when you want to place a camera and cast inside a spatial environment before generating footage.

1. Switch to 3D workspace. Choose a splat environment from the environment control or use its import tools for a supported asset.
2. Adjust the scene’s position, rotation, and scale. If a scan is upside down, use the flip control. Choose a skybox or upload a panorama when needed, then adjust its rotation.
3. Switch to Camera mode with V. Drag to orbit and Shift-drag to pan; frame the shot before adding camera keyframes.
4. Switch to Objects with W to position characters. Use the mode buttons if a shortcut is unavailable or a text field has focus.
5. Save the scene layout. Use saved scenes to revisit a setup, rename it, or duplicate it as a variant.

Splat environments and streamed scenes can have different controls; a disabled control is not available for that environment. Large environments depend on your browser’s graphics capability and may take time to load.

A character billboard is a flat visual reference positioned inside the 3D scene. Use it to block composition and establish the character’s intended look.

1. Drag a saved character into the viewport. Select the character in Objects mode.
2. Adjust position, rotation, scale, width, and height. Mirror the billboard when it suits the composition, checking accessories and screen direction afterward.
3. Choose the identity-and-outfit sheet for this setup. If the control shows Needs stance, generate a new transparent stance to update the visible billboard.
4. Reuse a previous stance from the stance library or generate one for the current pose. Inspect the silhouette, outfit, and face before using it in a shot.
5. Review the composition from the actual camera and save the scene. Describe the desired lighting in the shot direction or project style; generation uses the captured setup to guide a coherent treatment of the scene and cast.

Removing a character from a scene is different from deleting the saved character from your library. New footage is generated from the direction and references; the billboard is not a fully rigged animated 3D actor.

Captured boundary frames communicate the composition at the beginning and end of a generated shot.

1. Arrange the environment, cast, and camera for the opening. Select Capture Key to save the start keyframe.
2. Move the camera to the intended final composition and select Set Final Frame. Use Update Final Frame to replace that endpoint. Keep changes plausible for an 8-second segment.
3. Review the captured frames. Editor handles are hidden in captures, so judge the actual composition rather than the viewport controls.
4. Add action notes between keys and choose Preview Move to inspect the setup. Select Review & Generate to review the shot interpretation, action, dialogue, and generation settings. Keep identities, outfits, screen direction, and lighting consistent. Additional keys create additional 8-second segments, reflected in the estimate.
5. Generate after checking the cost, then review the rendered clip in project media. Save or duplicate the scene setup if you want another variant.

The model interprets the boundary frames. It does not promise an exact simulation of the camera path or object motion between them.

The media bin holds source assets. Adding one to the timeline creates an editable instance; the source remains available for other uses.

1. Add supported video, audio, or image files through the Media upload area. Wait for import or upload to complete.
2. Filter with All, Video, Audio, Image, Text, or FX. Generated clips and voice outputs also appear here.
3. Drag an asset to a suitable timeline track. Select the timeline instance to edit its placement and properties.
4. Use a card’s Rename control to give it a useful name. Where multiple versions exist, use its version chooser to inspect the version you need.
5. Use Download clip on a video card to save an individual source clip. This is separate from exporting your assembled timeline.

Deleting media is different from removing an instance from the timeline or the 2D sequence. Check whether an asset is still needed before using its delete action.

Free projects keep imported media and timeline edits on your device. This makes local editing possible without uploading all your source footage.

1. Keep your original files in a folder you control. Import them in the browser and device you intend to edit on.
2. Stay within the current local limits: 250 MB per file, 500 MB of media per project, and 100 media assets.
3. Export drafts regularly and wait for local saves to finish before closing the page.
4. If media is unavailable after changing browsers or devices, return to the original browser or re-add the source file where the app requests it.

Do not clear site data as a first troubleshooting step. Local media, unsynced drafts, and local timelines can be lost. Local editing also does not imply every feature works offline: sign-in, AI, stock searches, cloud saves, and cloud renders need a connection.

The library brings reusable footage and effects into the media bin.

1. Open Library from Media. Choose Effects, B-roll, GIFs, Anime CC, or VFX.
2. Search the footage library for the subject or action you need. Use VFX quick searches for common elements.
3. Select a result to inspect it, including its source and any licensing information shown. Confirm it is suitable for your intended use.
4. Import the clip or add the effect, then drag the resulting media asset into the timeline.
5. For an effect, preview it on footage and adjust it in the Inspector. Use the effect builder to create or edit a reusable custom effect when needed.

Availability depends on the connected source. A library label is not a blanket license for every item; retain attribution when the source requires it.

Use direct generation when you need a new shot while editing. Use the 2D workspace when you want to grow a scene through successive continuations.

1. Choose AI Clip in Media to open Generate AI Clip, then describe the shot.
2. Choose the offered model, resolution, duration, and aspect ratio. Add a character or visual reference where offered. Higher-resolution and verified-avatar generations use 8 seconds; unavailable duration buttons are disabled.
3. Use the negative prompt for unwanted elements such as overlays or extra people. Apply project style prompts if they support this shot.
4. If adding dialogue, choose a character with an assigned voice and review the voice options and combined cost.
5. Generate and wait for the new asset in Media. Preview it, then drag it onto the timeline.

Costs vary with model, resolution, duration, and voice work. The number shown before submission is the estimate for that configuration; 12 tokens per second is not the rate for every video setting.

Talking Head uses your verified avatar and character voice for a scripted on-camera clip.

1. Complete verified-avatar setup, then choose the person icon in Media, labeled Create talking-head clip, to open Talking Head. Select the avatar and intended wardrobe.
2. Choose a performance mode: Speaking for a short delivery, Timed cut for timed talking points, Idle loop for reusable idle footage, or Still hold for a held image.
3. Enter the dialogue or talking points and their timing. Use the visual prompt to describe framing and lighting. Short segments are easier to fit naturally than a dense script.
4. Review voice settings and the total cost. A timed cut may need multiple source renders; the displayed estimate accounts for the selected program.
5. Generate, then inspect lip movement, pauses, transitions, and audio. Assemble the finished segment with your other footage in the Editor.

Timed-cut drafts are kept in the current browser. Keep a copy of important script text. Verification and an appropriate assigned voice are prerequisites for the workflows that request them.

AI extension creates more footage; ordinary trimming changes which part of existing footage is used.

1. Hold Shift while dragging a video clip’s right edge. Drag into empty timeline space to request an extension, or toward the next clip to prepare a transition.
2. Review the proposed operation and give the next action clear direction. Extension uses source footage and endpoint context to guide continuity.
3. For a transition, choose the offered method. Insert bridge generates between boundary frames and shifts later clips; Overlay cut layers the generated transition over the cut.
4. Check the duration, model options, and token estimate before generating.
5. Preview both boundaries when the result arrives. Check image continuity and sound, then adjust the edit if needed.

An AI transition is separate from a regular editorial cut or built-in transition. Keep the original clips so you can compare results or choose a different approach.

The timeline is where the final scene is assembled. Tracks let picture, sound, text, and effects overlap.

1. Drag media onto a track and move it to the desired start time. Zoom the timeline for more precise placement.
2. Drag a clip edge to trim it. Images can be lengthened beyond their initial duration; video is limited by its source unless you generate an extension.
3. Select a clip, place the playhead at a cut point, and use Split selected clip. Remove or move the resulting pieces as needed.
4. Select multiple clips to move or group them together. Use copy and paste for repeated material; Inspector property-copy controls can reuse a treatment without duplicating the source.
5. Lock a track to protect it, mute it to silence its audio, or hide it to exclude its picture from preview. Check these states if something seems missing.
6. Preview the whole scene and use Undo or Redo to revise recent edits. Deleting a selected timeline item removes it from the edit, not from your source folder.

Scene timelines are separate. Check the scene picker before placing new footage or exporting.

Use the Inspector to fit footage to the scene and build layered compositions.

1. Select a visual clip and open Transform. Adjust its position, scale, rotation, and opacity, or use the preview controls for placement.
2. Set the scene’s canvas framing and use crop controls to remove unwanted edges. Check a vertical or square cut at its actual output ratio.
3. Open Blend to change how the clip combines with layers underneath it. Start with a normal/default blend, then compare alternatives such as Screen or Multiply.
4. Open Speed to change playback rate and the offered pitch-preservation settings. Recheck the resulting duration and audio alignment.
5. Play across the clip’s boundaries to catch empty frames, unexpected overlaps, and sound changes.

Preview volume and clip volume serve different purposes: preview volume is your listening level; the clip’s Audio settings affect the mix.

A keyframe stores a property’s value at a moment. Several keyframes can create a move, fade, or changing effect.

1. Select a clip and put the playhead where the change should begin.
2. Use the property’s keyframe control and set its starting value, such as position, scale, opacity, or volume.
3. Move the playhead later and set the next value. Add further points only where the motion needs to change.
4. Use the keyframe navigation controls to return to a point. Adjust easing where available to change how smoothly it accelerates or settles.
5. Preview the whole move at playback speed. Check the first and last frames as well as the middle.

Not every control supports animation. Use the keyframe controls shown for the selected property. Local draft exports approximate some animated transforms, opacity, and blends; inspect a cloud render for final delivery.

Masks decide which parts of a layer remain visible. Put another visual layer underneath when you want to replace the hidden area.

1. Select a video clip and open Remove background. Use the person mask to keep the character or keep the background.
2. Preview the result and adjust threshold and feathering to improve the edge.
3. Use brush corrections for areas the automatic mask missed. Undo or clear corrections if they make the result worse.
4. Use shape masks or crop controls when a geometric reveal is more appropriate. Adjust the shape and its placement, and keyframe supported properties when the mask must move.
5. Scrub the full shot, especially hair, hands, occlusions, and fast movement. A clean mask on one frame may not hold throughout the clip.

Automatic person masking works best with a distinguishable subject. Complex backgrounds and motion can require refinement.

Combine effects with compositing and audio to create a treatment you can reuse.

1. Open Effects in the media library, preview a preset, and add it to the project. Apply it in the timeline or through the selected clip’s effect controls.
2. Choose the target where supported: full frame, subject, background, or edges.
3. Adjust the effect parameters, blend mode, and wet/dry mix. Reorder or disable effects to understand how each contributes.
4. For an audio-reactive effect, select an audio band or modulation source and set its response. Preview with the intended soundtrack playing.
5. Open the effect builder to author a reusable preset. Save it, test it on actual footage, and use its edit control to refine it later.

Use a short representative segment to judge a complex effect stack before applying it across a long cut. A subject-targeted effect depends on the quality of the subject mask.

Enhance groups tools that analyze or process a source clip. Read the access requirement and cost shown for each operation.

1. Select the clip and open Enhance. Use Auto reframe to target a new ratio and adjust responsiveness, or Camera tracking to follow the intended subject.
2. Use Stabilize for camera shake, Reduce noise for noisy footage, Optical flow for frame interpolation, or Motion blur to adjust motion appearance. These cloud-processing tools require eligible access.
3. Use Enhance quality for the offered AI enhancement and output settings. This is separate from the non-AI cloud-processing allowance.
4. Run one operation and wait for completion. Inspect the processed result before choosing Use on this clip where offered.
5. Compare motion, fine texture, edges, and audio with the original. Reframing and enhancement can change the image in ways a single thumbnail will not reveal.

Cloud jobs need a network connection. If an action fails, preserve its exact message and check the job state before submitting the same request again.

Retouch processes footage on your device without credits and keeps the original in Media. It creates a corrected version for the selected part of the clip.

1. Under Retouch, choose Eye contact or Remove watermark to open Retouch clip. Preview the source. For eye contact, include a moment when the person looks at the camera.
2. Choose Find automatically or seek to a suitable frame and choose Use this moment. Adjust Correction strength, then check blinks and head turns with Show original.
3. For Watermark removal, draw a box around the mark with a small clean margin. Choose Fill from edges for a plain background or Copy nearby area and position a clean sample for texture.
4. For a moving mark, seek forward and add another position. Add points at direction changes; use Hold until next position for a mark that jumps.
5. Choose Apply to clip and keep the tab open until processing and saving finish. Long footage is processed in sections and saved as one result, up to 1080p at the source frame rate.

Eye contact is best suited to one person in a continuous, front-facing shot with steady lighting. Mark repair cannot recover hidden detail exactly. Inspect the moving result before using it in your final cut.

Standalone audio and embedded video sound can both be mixed in the Editor.

1. Add music, voiceover, or sound effects to Media and place them on the timeline. Listen to each source before balancing them together.
2. Select the audio or video clip and set Volume in decibels. Use Fade in and Fade out to soften its boundaries.
3. Add supported volume keyframes to lower music under dialogue and restore it afterward.
4. Use track mute to isolate parts of the mix. Check the preview listening volume if playback is unexpectedly quiet.
5. Play through every cut and export a short sample to check that speech remains clear and nothing clips or drops out.

For character-specific dialogue, assign a voice in Character Creator first, then generate voice from the storyboard or use the Editor’s character-voice tools.

Markers give you timing targets for cuts, titles, and effect changes.

1. Select audio and use Beat markers to analyze its rhythm.
2. In the timeline’s beat-marker control, choose High / hi-hat, Low / bass, Both beat bands, or No beat markers.
3. Place the playhead at a custom cue and select Add marker at playhead or press M when the timeline has focus.
4. Use snapping to align clip boundaries with visible timing targets. Shift-click markers to add them to the selection; double-click a marker to remove it.
5. Listen to the result. Automatically detected transients are editing aids, so choose the beats that fit the story rather than cutting at every marker.

Hiding beat markers also disables snapping to those beats. Custom markers are useful for dialogue cues, chapter starts, and review notes.

Text is a reusable media asset that you can place over footage on its own track.

1. Create a text asset in Media and drag it onto the timeline at the moment it should appear.
2. Select it and open Typography. Enter the text and adjust font, size, alignment, color, and spacing using the available controls.
3. Browse the Google Fonts picker and preview the font with your own words before selecting it.
4. Adjust transform, opacity, outline, shadow, blend, and animation controls. Use keyframes or clipping-mask controls for a reveal when needed.
5. Trim the text clip to its intended duration. Check readability at the final frame size, including on a phone for vertical video.

Keep important text inside the visible frame. Fonts and effects that look good in a large preview still need a check in the exported video.

Turn speech or lyrics into timed text, then proofread the words and timing before delivery.

1. Select the audio source or a video with audio and open Lyrics & captions.
2. Run the offered transcription action after checking its cost. Choose Phrase timing or Per word to suit the presentation.
3. Review every phrase. Correct names, punctuation, and misheard words, and adjust start and end times.
4. In word mode, refine individual word timing where necessary. Style the resulting text with the typography and animation controls.
5. Preview captions against the audio and export a sample. Check fast speech, pauses, overlapping voices, and the final frame.

These are timeline text layers for your edit. They are separate from the optional captions on CineGen’s landing-page walkthrough, which viewers can enable in that video’s player controls.

Export the current scene’s timeline after checking the selected scene, frame, visible tracks, and audio mix.

1. Choose Export and enter a Filename.
2. For a quick local check, choose the draft export type and quality, then Export draft. Keep the tab open while it renders and downloads.
3. For a final render, choose the cloud export type and a preset, then Queue cloud render. Options include YouTube 1080p, 2K QHD, 4K UHD, vertical TikTok / Reels, and ProRes 422 HQ for a MOV master. This requires eligible paid access.
4. Open Exports to follow the job from queued to ready. Download the completed result; check the error details and retry option if it fails.
5. Play the downloaded file from beginning to end. Check timing, caption placement, transitions, effects, picture quality, and sound before sharing.

Free draft exports are 720p and include a CineGen preview watermark. Local drafts approximate some keyframed transforms, opacity, and blending; use a cloud render for the final cut. Download completed exports promptly because retained export files can expire.

Teams organize collaborators and can share a token pool. Team membership and access to a specific project are separate choices.

1. Open Team and create or select the team available to your account.
2. Invite an existing CineGen user by handle or email. Have a new collaborator create their account first if the invitation cannot find them.
3. Review membership and roles. Use Share & Review → Project access to grant the appropriate people access to an individual project.
4. Open Team usage to inspect pooled token use, seats, and billing information available to your role.
5. Have the team owner manage seat or billing changes, and verify the selected billing context before purchasing tokens.

An invitation is not a reason to share your login. Each collaborator should use their own account so permissions and activity remain attributable.

Give each collaborator the access needed for their work.

1. Open Share & Review for the project and find Project access.
2. Choose a team member or enter their handle or email through the available invitation controls.
3. Assign Editor for people who should change project content or Viewer for people who should watch and comment. Project owners manage access.
4. Ask the collaborator to open Shared with me on their Dashboard or use the project link.
5. Change or remove membership when it is no longer needed. Review link access separately, since links are another way to reach a scene.

If someone cannot open a project, check the signed-in account and project membership before assuming the project is missing. A viewer’s read-only controls are intentional.

Review links let someone watch a selected scene without giving them your entire production workspace.

1. Select the Scene and add an optional label, such as “Client review, cut 3.”
2. Choose Read only, Comment, or Edit (sign-in required). Comment links allow timestamped feedback without requiring an account; edit links also grant signed-in project membership.
3. Set Expires in to 24 hours, 3 days, 7 days, or 30 days, then choose Create link.
4. Open the review page yourself to check the scene and access level, then copy the link to share it.
5. Track feedback in Review comments. Revoke the link when review is finished or the audience changes.

Anyone holding an active link can use the access it grants. If an edit link has already granted membership, review that person’s Project access as well as revoking the link.

A useful review note identifies the moment or position that needs attention and describes a concrete change.

1. Right-click a timeline location or a supported scene canvas to place a contextual comment. External reviewers can leave timestamped notes through a comment-enabled review link.
2. Write the requested change and use @ to mention a collaborator where offered.
3. Open a pin to read its thread and reply. Use Share & Review → Review comments to see feedback together.
4. Select a comment’s context link to return to the relevant time or scene position.
5. Use Hide pins or Shift + C where supported to clear your own view while editing. This does not hide the feedback from collaborators.

Check the notification feed for mentions and replies. Notification preferences control what you see; they do not delete the underlying conversation.

Use the project’s automation controls to connect review activity with your delivery workflow.

1. Open Integrations & automation with permission to manage the project.
2. Enter a Slack incoming-webhook URL if you want review and render events sent to your chosen channel, then save the settings.
3. Configure the available approval/render option if approval should queue a render automatically. Review the settings before enabling it.
4. Use the project’s review controls to record the cut’s decision and check the resulting render job in Exports.
5. Inspect the Audit log for recent access, membership, review, and render events when you need to understand what happened.

Treat a webhook URL as a credential and only use a destination you control. Automatic rendering still depends on the project’s export eligibility and a valid renderable timeline.

Tokens pay for AI actions. Check the estimate immediately before submitting, especially after changing video settings or adding a voice.

Other video models, resolutions, timed programs, voice changes, and enhancement actions have different rates. The app’s estimate is the source for the action you are running; these examples are not a universal rate card.

1. Check your balance and whether the action uses your personal or team allowance.
2. Review the estimate, then submit once. Wait for the job state before trying again.
3. If generation fails, read the exact error. Failed billable generation jobs are designed to refund the debit automatically; check the balance and usage record.
4. If the refund or result is missing, contact Support with the project, action, time, job details, and displayed charge.

An unwanted but successfully generated result is different from a failed job. Removing a generated asset or sequence entry does not itself refund tokens.

Manage the subscription associated with your account, or the team billing context you are authorized to manage.

1. Open Settings → Billing to see the plan, balance, and billing controls.
2. Compare available plans and billing intervals before upgrading or changing your subscription. Review the checkout’s price and renewal terms.
3. Use the offered token packs or custom top-up when you need more tokens. Confirm the destination account or team before completing checkout.
4. Use the billing-management portal to manage payment details, invoices, or cancellation options.
5. Return to CineGen and verify that the plan or balance updated. Contact Support with the receipt if a completed payment is not reflected.

For current prices and included allowances, use Pricing. Billing questions, refunds, and subscription corrections need human support; do not include full payment-card details in a ticket.

Keep your profile recognizable for collaboration and protect the account that owns your projects.

1. Update your display name, profile photo, and unique CineGen handle in profile settings, then save. Your handle helps teammates find and mention you.
2. Use the password settings for a password-based account. If you use social sign-in, use the sign-in method associated with that account.
3. Set up Two-factor authentication with a compatible authenticator app: scan the QR code and enter the six-digit code to verify setup.
4. Use Email two-step verification if you want the offered email-code protection. Complete the email challenge to enable or disable it.
5. If a security change requires a recent login, preserve your work, sign in again, and retry the change.

Use the account-deletion settings only after reviewing their confirmation and scheduling details and keeping any work you need. For account access, ownership, or privacy requests, contact Support. Never send passwords, authentication codes, or setup secrets in a ticket.

Choose which activity should interrupt you and which updates should arrive by email.

1. Open notification settings and review the In-app feed categories.
2. Turn individual categories on or off to control what appears behind the notification bell.
3. Under Email, choose whether to receive the Weekly account digest and Product updates.
4. Open the bell to follow enabled activity back to the relevant comment, mention, or account event.

Turning a feed category off hides it from your feed without deleting its history. Optional product updates can also be unsubscribed from through their email controls.

Use shortcuts when the relevant workspace has focus. Text fields and dialogs may handle the same keys differently.

For touch input or when a shortcut does not apply, use the corresponding on-screen controls. The guide’s chapter links can be bookmarked or shared directly.

Start with the specific failure and preserve your work before changing browser settings.

1. Confirm the latest save or export and keep a copy of unsaved prompt/script text.
2. Retry the specific action once when appropriate. If it still fails, refresh after preserving work.
3. For browser-processing issues, keep the tab active and test a shorter source clip. Check whether the feature reports an unsupported browser capability.
4. Report repeated errors with enough detail to reproduce them. Avoid clearing all site data, since local work may be stored there.

AI results can vary even when a request succeeds. For identity drift or ignored framing, include the intended direction, selected references, and the point in the clip where the result diverges.

Open Support while signed in to start or continue a support conversation. If you cannot access your account, email support@aionx.aionapp.org.

1. Describe what you tried to do and what happened instead. Include the exact error text.
2. Add the project and scene, relevant clip or job, approximate time, browser, and device. Include a screenshot or a short reproduction when useful.
3. Explain which steps you already tried and whether the problem survives reopening the project.
4. Follow replies in your support conversation. Billing, ownership, verification, privacy, and data-loss cases require human assistance.

For additional workflow articles, visit CineGen’s blog. For account terms and data handling, read the Terms of Service and Privacy Policy.

Keep credentials and private payment details out of support messages. Share only the project material needed to understand the issue.