What Frame Highlights Does Not Do
Frame Highlights keeps files in the workspace after accept, select, and save. It does not publish to external platforms, and it does not turn adoption, queue, or running counts into views or payments.
Team assets exist only after review
The boundary of Frame Highlights fits in one line: only after review does something become a team asset. A recommended clip is a 待确认片段 until it is accepted. The workbench says 还有待确认片段，可逐个采用或驳回. Covers on the task detail are 已选定 or 已取消 for 16:9, 1:1, and 9:16. Only after those decisions, and after 保存为素材 or a batch save by task, do full videos, highlight clips, cover images, and subtitle files settle in the asset library. The product page is https://ajuno.xibalans.com/image-video-processing/ .
When analysis ends and the task status becomes 已完成, that means this background run finished. It does not rewrite pending clips into enabled assets, and it does not rewrite an unselected cover into 已选定. 立即处理 opens the review. It does not skip review and write the library by itself.
Completion is a status on the dashboard and on the task, beside queued, analyzing, and failed. A completed task can still show pending clips. A completed task can still have a cover ratio marked 已取消. The asset total does not have to increase at the moment the run completes, because save is a separate action. Treating completion as storage, or storage as publication, skips two boundaries the product actually draws.
The workflow has no external publishing step
The overall path is upload video, choose a template, review results, and keep them for reuse. The upload page order is select a template, confirm the template, upload videos, and background analysis. Both stop inside the workspace. Frame Highlights does not publish to external platforms. Navigation is 工作台, 上传视频, 任务列表, 素材资产库, and 数据大屏. None of those is an external account or an external channel.
Accept, reject, 已选定, 已取消, save as an asset, 启用, 归档, and filtering by 合集 are all internal actions. Same-source compare at /tasks/compare?a=&b= or ?ids= compares parameters and aligned segment differences. It is not a publishing order. At least two tasks are required. 克隆重跑 and the button 对比同源任务 are only the ways to open that compare.
There is no step after the library that says the file has been handed to another product. Recent tasks, the latest five, and recent assets remain lists inside the workbench. The page size of 24 and the collection filter remain library controls. The Top 15 tags remain a dashboard list. None of those surfaces is described as an outbox.
Numbers on screen are a snapshot of this workspace
The workbench figures 素材总数, 待确认片段, 精彩采纳率, and 封面采纳率, and the dashboard cards for 近 7 天, 近 15 天, or 近 30 天 — 任务总数, 已完成, 素材总数, 复用率, 精彩采纳率, and 封面采纳率 — are a snapshot of the current workspace when the page is open. The same is true of 处理趋势 with 任务完成, 片段产出, and 素材新增, of 热门标签 Top 15, and of 资产构成 for 精彩片段, 封面, and 素材. They describe the window you are looking at.
Those numbers should not be written as a score shared by every team, and they should not be written as views. The default preset weights 运动强度 1.00, 音频能量 0.65, 人脸出现 0.35, 镜头切换 0.55, and 画质优秀 0.45 are parameters of 人物动感高能. They are not adoption rates, and they should not be copied onto the other four presets. The header queue count and running count are a snapshot of this moment too. 系统运行正常 describes the status at that time. It does not guarantee the length of the queue an hour later.
A sample seen on one visit can be reported only with that limit attached: this workspace, this range, this moment. Changing the range from seven days to thirty days is allowed to change the cards. Saving or archiving assets is allowed to change the asset total. Accepting a pending clip is allowed to change highlight adoption. None of those movements creates a benchmark for a different team or a different week.
Queue and running counts are not a delivery guarantee
The queue count and the running count answer how much is waiting and how much is running now. 任务状态分布 for 排队中, 分析中, 已完成, and 失败 answers how those statuses are mixed across the selected days. Neither one is a delivery contract, and neither one is a scheduled broadcast. Failed means this run did not finish. The task itself has to be inspected. A failed status does not imply a settlement.
Upload limits should not be read as promises either. Formats are avi, mkv, mov, and mp4. Each file may be at most 2048 MB. Files cannot be selected before the template is confirmed. 封面高清增强 is only a checkbox. Meeting those conditions only means the file can enter background analysis under the confirmed template. Analysis can still fail. Recommendations still need review. Covers still need a per-ratio decision.
The five presets remain strategies, not service levels. Confirming 画质优秀 does not promise a sharp frame in every second. Confirming 综合动感高能 does not promise that fast motion will be found. The preset text says what the strategy prefers. The run can complete, fail, or return recommendations that the team rejects. Queue depth does not change that review duty.
Publishing, views, and settlement stay outside
What Frame Highlights does not do is treat a reviewed result in this workspace as already published, already watched, or already paid. The six landing capabilities are batch video upload, analysis templates, highlight recommendations from five signals, covers in three ratios, the asset library, and task and data review. External publishing is not among them. View statistics are not among them. Payments are not among them.
An MCN team can look up enabled assets 24 to a page, group them with a collection, and use the dashboard to see adoption over a chosen number of days. Those capabilities help review and reuse. If a channel post, a view result, or a settlement is required, it happens outside Frame Highlights. Copying one rate off the screen into a plan and saying every project will match it goes past what this product actually shows.
The tagline area calls the product AI 视频内容工作台. A workspace for analysis is the right size for that sentence. Templates, signals, covers, review, the library, and the dashboard fill the workspace. They do not add a publisher, a view counter, or a payout. Keeping that list exact is what makes the nine-part guide usable: each screen does the job its labels describe, and no screen quietly finishes a job the product does not have.
Does a high highlight adoption rate mean there will be views?
No. 精彩采纳率 is a snapshot of the current workspace. It describes clip adoption. It does not describe watch counts, and it is not a universal standard.
After an asset enters the library, is it sent out automatically?
No. 保存为素材 only puts the result in the asset library. Frame Highlights does not publish to external platforms. Archive and enable do not trigger an external post either.