Skip to main content

Portal Tools

Valossa Portal provides a suite of web-based tools for working with video analysis results without writing code. These tools complement the API by offering visual interfaces for exploration, editing, and search.

Valossa Report​

Valossa Report is a visualization tool that presents your video analysis results in an interactive, human-readable format.

Overview Section​

The Overview provides a quick summary of the most prominent detections in the video. Tags representing detected concepts are displayed visually, and you can:

  • Browse through detections using navigation arrows
  • Search for specific concepts within the video using the magnifying glass
  • Click on tags to jump to the relevant video position

Heatmap Section​

The Heatmap displays a timeline of the video with detections placed along it. Key features:

  • Each detection type has its own colored row (faces, visual context, audio context, etc.)
  • Detections are grouped by type and sorted by prominence within each group
  • Color depth indicates detection density and confidence at each time position
  • Click on any colored cell to play the video from that timecode
  • Adjustable timeline resolution and detection count
  • Configurable confidence thresholds for filtering

Tag & Train​

The Tag & Train tool allows you to:

  • Edit names and genders of detected faces
  • Save changes to the metadata
  • Train the AI to recognize the corrected face identities in future analyses
  • Index edited results into Valossa Search automatically

Access Tag & Train by clicking the "Tag & Train" button above the face detections in a Valossa Report, or the pencil icon next to a person name.

Accessing Valossa Report​

On the Valossa Portal home page, each successfully analyzed video has a link to its Valossa Report. Demo reports are available by clicking "Demos" from the home page (requires login).

Captions Editor​

The Captions Editor is a visual tool for reviewing and correcting speech-to-text results to produce accurate subtitles.

Features​

  • Text editing: Modify the content of any caption segment
  • Timing control: Edit start and end times for each segment
  • Segment management: Add or remove caption segments
  • Export: Download corrected captions in SRT or VTT format

Workflow​

  1. Analyze a video with speech-to-text enabled.
  2. Open the Captions Editor from the video's results page on the Portal home page.
  3. Review and correct any transcription errors.
  4. Adjust caption timing as needed.
  5. Export the final captions in your preferred format.

The Captions Editor is particularly valuable when you need professional-quality subtitles. Valossa AI provides the initial transcript, and the editor lets your team refine it efficiently.

Valossa Search enables searching across all your analyzed videos and within individual videos.

Capabilities​

  • Search across all analyzed videos on your account
  • Search inside a specific video from within its Valossa Report
  • Targets: visual concepts, topics, categories, audio sounds, speech and transcript results, titles, and descriptions

Search Examples​

  • "car" -- Find all scenes with cars across your video library
  • "Arnold Schwarzenegger with a gun" -- Find specific person-object combinations
  • "music" -- Find segments with detected music

Advanced search supports:

  • Detection type filtering: Limit results to specific detection types
  • API key filtering: Search only jobs analyzed with a specific API key
  • Title filtering: Search within jobs matching a title string
  • Cross-video search: Use the search field on the Valossa Portal home page
  • Within-video search: Use the search field within a specific Valossa Report
note

Valossa Search only returns results from videos that have been analyzed on your account.