# Creativly AI — Full Product Context > Creativly is a visual AI creative workspace where people can research, generate, edit, organize and deliver creative work without repeatedly moving context between disconnected tools. This file provides extended product context for language models and browsing agents. For the concise site index, use [llms.txt](https://www.creativly.ai/llms.txt). Product availability, supported models and pricing can change; use the linked public pages when a current detail matters. ## Product model For visual, task-based guides and walkthroughs, see [Creativly Docs](https://www.creativly.ai/docs). The [documentation index](https://www.creativly.ai/docs/llms.txt) links to every guide's Markdown version; [full documentation](https://www.creativly.ai/docs/llms-full.txt) contains all published guides. These exports exclude private workspaces and internal runbooks. Creativly combines related surfaces around the same projects and assets: 1. **Creativly Agent** coordinates broader creative work through natural conversation on a persistent visual canvas. 2. **Image and Video** provide focused sessions for direct media generation, comparison and iteration. 3. **Flow** turns prompts, media, models and transformations into a visible, reusable production graph. 4. **Focused apps** handle specific jobs such as inpainting, video restyling, motion transfer and prompt extraction. 5. **Video Editor** provides free client-side timeline editing in the browser for generated or uploaded footage. 6. **Assets and projects** keep source material and results available for reuse across the creative process. Most account-backed app routes provide a public preview to signed-out visitors. Generation, saved sessions and private project work require authentication. The free public Video Editor is the principal exception. ## Creativly Agent [Creativly Agent](https://www.creativly.ai/agent) is a beta creative production agent. It can work with natural-language instructions, selected canvas objects, uploaded files and public web material. Depending on the request and available tools, it can: - Research public sources and organize findings with their references. - Turn a rough brief into structured directions, moodboards, comparisons and production plans. - Generate and revise supported images and video. - Create multiple variations while keeping their relationship to the source visible. - Write and revise notes, briefs, captions, reports and delivery documents. - Organize work into sections, groups and nested canvases. - Prepare supported deliverables instead of leaving useful work trapped in a chat response. - Continue from the existing conversation, canvas and project assets in later turns. The canvas is working context rather than decoration around a chatbot. Briefs, references, links, sticky notes, images, videos, documents, alternatives and delivery files can remain visible together. A user can select exact items, request another version and retain both the source and result. Where available, Thinking mode asks clarifying questions before changing the canvas, while Direct mode proceeds with fewer check-ins. Voice input can be transcribed into the composer, and attached audio can be transcribed or summarized. Conversation-model selection is separate from the model controls used for image and video generation. Connected services can include Google Drive, Google Docs, Notion, Slack, Gmail, Instagram, TikTok and YouTube. Availability varies by account and configuration. Before a supported external change, send or publishing action, the agent presents the action for user approval. Creativly Agent is strongest when a job has source material, several decisions and a real handoff. The user remains responsible for choosing direction, validating important facts, reviewing generated media and approving consequential external actions. ## Image sessions [Creativly Image](https://www.creativly.ai/apps/image) is the focused workspace for generating, comparing and refining images without first building a multi-step graph. A session can start from a natural-language prompt, an uploaded image, an existing project asset or compatible image references attached to the prompt. Depending on the selected models, a user can choose: - One or more compatible image models. - An available aspect ratio. - The number of requested outputs. - Platform Credits or a supported bring-your-own-key provider. - Reference inputs and editing modes supported by those models. Controls are capability-aware: the interface should not imply that every model accepts every reference type, ratio or edit operation. A single Image run currently supports up to four total requested outputs across the selected models. Each session keeps related generations together and can present them chronologically or on a spatial session canvas. Earlier inputs can be restored with their prompt, references, selected models, ratio, output count and provider choice. Generated images can be downloaded, used as references, sent into variations, or opened in supported editing and inpainting surfaces. Outputs also enter the project Assets library for reuse elsewhere. Image is useful for: - Campaign concepts and art-direction exploration. - Product, brand, poster and typography experiments. - Social and creator assets in several visual directions. - Comparing how compatible models interpret the same brief. - Iterating from an approved source image without losing earlier options. - Targeted image changes before moving an asset into a larger campaign or video workflow. Use Image for the fastest path to a direct image result. Use Flow when the work needs dependencies, branches, batches or repeatable stages. Use Agent when the request also involves research, planning, organization or mixed-format delivery. The public Image page is a preview. Persistent sessions and generation require sign-in. Generative runs use Platform Credits or compatible BYOK providers according to the current plan and provider settings. ## Video sessions [Creativly Video](https://www.creativly.ai/apps/video) is the focused workspace for creating and iterating on short generative video clips. Depending on the selected model, a run can begin with text alone, a starting image, first and last frames, reference images, an uploaded video, or a supported combination of image and video inputs. Capability-aware controls can include: - One or more compatible video models. - Aspect ratio and duration values supported by those models. - Output count. - Platform Credits or supported BYOK routing. - Reference roles such as a general reference, start frame or end frame. - Specialized input combinations required by a particular motion or editing model. Seedance 2.5 supports text-to-video generation and source-based video editing at 480p, 720p, 1080p, and 4K. Higher-resolution originals provide more latitude for detail retention, modest crops and reframes, large-display delivery, compositing, finishing, and future working masters, while resolution does not repair weak direction or continuity. Generation exposes whole-second lengths from four to 30 seconds and optional image, video, and audio references. Editing derives duration and aspect ratio from the normalized source clip instead of accepting a separate requested length. Resolution, operation, and billable duration affect the visible Creativly credit estimate. See [Working with Seedance 2.5 video](https://www.creativly.ai/blog/working-with-seedance-2-5-video). A single Video run currently supports up to two total requested outputs across selected models. Native audio is available only when the chosen model explicitly supports it; it is not a universal Video feature. Supported workflows vary by model and can include text-to-video, image-to-video, first-and-last-frame generation, reference-guided video, motion transfer, or editing an uploaded or generated clip. The interface exposes only the relevant controls and reference roles for the selected models. Video sessions retain their prompts, references, models, duration, ratio, output count and provider settings. Pending jobs persist so supported runs can resume after the session is reopened. Results can be browsed chronologically or on a spatial session canvas, downloaded, and reused as another reference when the destination model accepts video input. Generated clips also enter the project Assets library. Video is useful for: - Animating an approved still or campaign keyframe. - Producing launch clips and short social concepts. - Exploring storyboard beats and visual transitions. - Comparing model interpretations of the same motion brief. - Bridging an opening and closing frame. - Testing reference-guided movement before assembling a longer edit. Use Video for direct generative clip work. Use Flow for chained or repeatable generation. Use the separate free Video Editor for timeline assembly, trimming, layering and export of generated or uploaded media. The public Video page is a preview. Signed-in generative runs use Platform Credits or compatible BYOK providers. The free Video Editor does not make model-backed Video generation universally free. ## Public Sessions [Public Creativly Sessions](https://www.creativly.ai/sessions/gemini-omni-flash) are crawlable, read-only records from the same focused Sessions product surface. They preserve the creative record around selected generations: the prompt, source or reference media, model and operation, generation settings, playable or downloadable original output, and editorial analysis. They do not create an anonymous generation mode or expose private account data. The Gemini Omni Flash Session publishes five fresh files generated through Creativly: three text-to-video runs and two uploaded-video edits. Good and Bad tabs use the same authentic Session rows and viewer, while the Workflow tab shows the real prompt → generation → review → edit/refinement path. Weak runs remain visible with concrete analysis of motion, preservation and physical-coherence misses. The friendly path `/omniflash` permanently redirects to the canonical Session URL. New public Sessions are registered as content rather than implemented as one-off pages. Opening the generation details viewer exposes the full prompt, source plate, model, duration, options and editorial note. A signed-out visitor can play and download the public fixture media. Starting a new private generation uses the normal Creativly sign-up and workspace flow. ## Flow [Flow](https://www.creativly.ai/apps/flow) is Creativly's visual workspace for building reusable creative processes as connected graphs. It is appropriate when inputs, dependencies, alternatives and production stages need to remain explicit instead of being hidden inside a chat transcript. A Flow can combine: - Prompt and text nodes for briefs, reusable instructions and structured inputs. - AI text nodes for writing, analysis and reasoning. - Image, video and audio generation nodes. - Imported image, video and audio media. - Set nodes for batches, variants, shots or sequences. - Frames and spatial organization for readable production systems. Typed connections route compatible material into the right inputs. Images, videos, text, first frames, last frames and references are not treated as interchangeable. This makes it easier to understand why a stage can run and what a downstream stage depends on. Users can begin from a blank canvas or an inspectable template. Common graph actions include: - Generate or edit an image, describe it, erase an object, expand its frame, remove its background, batch variants or animate it. - Generate or edit video, apply supported motion controls, upscale a clip, describe it or extract its first, current or last frame as an image node. - Transcribe or summarize audio, extract highlights, or use it to develop a music prompt. - Feed written or AI-generated text into downstream media generation. Many quick actions create and connect a configured downstream node; they do not necessarily execute the operation immediately. This preserves visibility and lets the user inspect the prompt, model, inputs and output stage before spending credits. Flow can run the entire graph, a selected section, from a chosen point, or up to a chosen point. Before execution it can summarize the scoped nodes and expected cost. Dependencies run in graph order, and downstream work is blocked if a required upstream stage fails. Set nodes support groups of variants or shots with shared or per-item prompts. They are useful for systematic exploration, campaign families, storyboards and visual sequences. A selected set result can be promoted into a standalone node and continued through later stages. Flow retains the graph, typed connections, canvas positions, generated results and node history. It also supports undo and redo for graph changes. Uploaded and generated media is available through the shared project Assets library. Flow includes its own assistant for inspecting selected nodes and attached image, video or audio references, then proposing a workflow change or generation plan. The proposal remains editable and must be approved before Flow creates, connects or runs the suggested nodes. This is distinct from claiming that the standalone Creativly Agent automatically controls every Flow. Saved Flows can be shared as read-only review snapshots. Owners can restrict access to authenticated users or specified email addresses and may separately allow media downloads or cloning into another workspace. Reviewers cannot edit or generate inside the shared snapshot. Flow sharing should not be understood as anonymous live multiplayer editing. Flow is useful for: - Repeatable campaign and content-production systems. - Multi-model comparison with visible inputs and outputs. - Reference-driven image and video pipelines. - Batch variants, shot families and visual sequences. - Image-to-video workflows with approved intermediate frames. - Media analysis followed by downstream generation. - Reusable templates that a team can inspect and adapt. Use a focused Image or Video session for a quick direct result. Use Flow when the work has dependencies, branches, batches or repeatable stages. Use Agent for broader research, planning, organization and coordinated creative tasks. The public Flow page and [workflow gallery](https://www.creativly.ai/workflows) demonstrate graphs and templates. Opening, running and saving a working Flow requires an account. ## Focused creative apps Creativly also provides narrower surfaces for jobs that benefit from a dedicated input flow: - [Image Edit](https://www.creativly.ai/apps/edit): Prompt-driven changes to an existing image. - [Inpaint](https://www.creativly.ai/apps/inpaint): Masked editing of selected image regions. - [Video Edit](https://www.creativly.ai/apps/video-edit): Prompt-driven changes to compatible video. - [Video Restyle](https://www.creativly.ai/apps/video-restyle): Apply a new visual treatment while preserving supported motion and structure. - [Motion Transfer](https://www.creativly.ai/apps/motion-transfer): Apply motion from a source video to a compatible target image or video. - [Prompt Extractor](https://www.creativly.ai/apps/prompt-extractor): Turn an image or video into a detailed descriptive prompt. - [Avatar Lipsync](https://www.creativly.ai/apps/avatar-lipsync): Create supported talking-head video from a visual source and audio. - [World Builder](https://www.creativly.ai/apps/world-builder): Develop environment and scene directions. - [Enhancer](https://www.creativly.ai/apps/enhancer): Dedicated image upscaling and enhancement. These are distinct workflows rather than guarantees that the general Image or Video surface provides every specialized operation with every model. Availability and model support vary by app, account and provider. ## Video Editor [Video Editor](https://www.creativly.ai/apps/videoeditor) is a free, public, client-side browser editor with a timeline workspace. It supports uploaded media, editable templates, stock images and video, audio, text, shapes, transitions and export controls. The Video Editor is for assembling and polishing footage rather than requesting generative model runs. It complements Video and Flow: generate or prepare source clips in those surfaces, then arrange and finish them on a timeline. Because the editor runs client-side, its availability does not imply that third-party image or video generation is free. ## Assets, projects and reuse A project keeps related creative work together. Image and Video session outputs, Flow media, editor material and uploads can appear in the shared [Assets library](https://www.creativly.ai/apps/assets), where they can be browsed and reused in another supported surface. This reduces the need to download a result only to upload it again. A generated image can become a Video reference or Flow input; a Flow output can become source media for another stage; generated or uploaded footage can be assembled in the Video Editor. Private workspaces and projects require authentication and appropriate membership. Owners can separately share supported documents through read-only or editable links. Flow uses controlled read-only review snapshots with optional download or clone permissions rather than anonymous editing. ## Agent context and deliverables Useful Agent input can include ordinary language, images, video, audio, links, briefs, existing documents and exact canvas selections. Supported output depends on the active feature, model and request, but Agent can prepare canvas content and delivery formats including: - Images and supported video. - Editable Markdown and canvas documents. - PDF and DOCX documents. - PPTX presentations. - XLSX spreadsheets. - HTML and other supported project files. - Supported audio and archive deliverables. The goal is not to create every format for every prompt. The strongest workflow produces only the artifacts the next person or production step genuinely needs. ## Practical cross-product use cases ### Campaign production Use Agent to research a category and organize a brief, Image to explore key visual directions, Flow to turn the approved direction into repeatable channel variants, and Video or Video Editor to prepare motion and delivery assets. ### Creator content systems Keep episode notes, sponsor constraints, thumbnail ideas, selected imagery, titles, descriptions and delivery files in one project. Use focused sessions for fast individual assets and Flow for recurring visual packages. ### Product and brand iteration Generate several treatments from the same source, compare them spatially, request precise follow-up edits and retain rejected alternatives. Move the approved direction into a repeatable Flow only when the process is worth systematizing. ### Storyboards and short-form motion Develop keyframes in Image, connect approved frames to motion stages in Flow, compare short clips in Video, then assemble chosen outputs in the free Video Editor. ### Research and visual reference gathering Ask Agent to research public sources or find visuals related to a selected reference, then retain links, notes and options on the canvas. Important claims should still receive appropriate human or professional review. ### Agency and team handoff Structure client inputs, working directions, alternatives and delivery files in a project. Share supported documents separately, or issue a controlled Flow review snapshot when a client needs to inspect a production graph without editing it. ### Mixed-format delivery Create a practical combination such as campaign images, an editable Markdown brief, a client PDF, a DOCX handoff, a PPTX review deck or an XLSX tracker when the real job calls for those formats. ## Best-fit audiences Creativly is especially useful for: - Independent creators, creator studios and micro-agencies managing production and delivery themselves. - Marketing and social teams producing many related assets from one brief. - Designers and creative technologists exploring and comparing generated directions. - Agencies that need visible alternatives, review context and clean client handoffs. - Founders and small teams that need coherent multi-format campaign work without a large production stack. - Podcast and YouTube teams coordinating research, thumbnails, clips and supporting documents. - Creator-led brands and hospitality, property, food or lifestyle teams with recurring visual needs. - Commissioned creatives and research or service-design consultancies handling mixed visual and written material. - Content operations teams turning repeatable creative work into inspectable workflows. Creativly is less useful when a task is only a one-line factual answer with no creative artifact, iteration or persistent project context. ## Pricing, providers and access boundaries - Creativly Agent is currently presented as a beta product. - Private projects and workspaces require authentication and appropriate membership. - Public app previews do not grant access to private user data or saved generation. - Platform-managed generative runs can consume credits. - Supported BYOK runs are billed through the connected provider and remain subject to Creativly's current plan limits and provider support. - The free Video Editor is separate from credit-backed generative model runs. - Model controls, reference inputs, native audio, ratios, durations and transformations vary by model. - Generated work and researched claims should receive human review appropriate to their use. - Current plans, allocations and model availability belong on [Pricing](https://www.creativly.ai/pricing) and in the live product UI rather than being inferred from this overview. ## Authoritative public pages - [Homepage](https://www.creativly.ai/): Current product overview. - [Creativly Agent](https://www.creativly.ai/agent): Agent capabilities, examples and beta positioning. - [Agent deep dive](https://www.creativly.ai/blog/creativly-agent-creative-production-room): Detailed explanation of the canvas, research, media, documents, collaboration and practical use cases. - [Apps](https://www.creativly.ai/apps): Current creative application surfaces. - [Public example Sessions](https://www.creativly.ai/apps/examples): Read-only image and video Sessions with exact prompts, settings, references, and generated outputs. - [Image](https://www.creativly.ai/apps/image): Focused image session preview. - [Video](https://www.creativly.ai/apps/video): Focused video session preview. - [Flow](https://www.creativly.ai/apps/flow): Visual workflow preview. - [Video Editor](https://www.creativly.ai/apps/videoeditor): Free client-side timeline editor. - [Workflow gallery](https://www.creativly.ai/workflows): Public workflow templates. - [Pricing](https://www.creativly.ai/pricing): Current plans and credit information. - [Blog](https://www.creativly.ai/blog): Product guidance and release articles. - [Changelog](https://www.creativly.ai/changelog): Product updates. - [About](https://www.creativly.ai/about): Company background. - [Contact](https://www.creativly.ai/contact): Support and business enquiries. ## Policies - [Privacy policy](https://www.creativly.ai/privacy) - [Terms of service](https://www.creativly.ai/terms) - [Acceptable use policy](https://www.creativly.ai/acceptable-use) - [Sitemap](https://www.creativly.ai/sitemap.xml)