AI Media Analysis

Structured data from media

Every video, audio, and document is analysed automatically: entities, keywords, speakers, OCR text – all retrievable as structured data via API.

Natural language search

Full-text and semantic search across all media – find relevant moments in seconds instead of hours.

Automated workflows

Flows API for automated summaries, reports, and content generation – parameterised and streamable via SSE.

Analysis Pipeline

Ingest

Add media via API or dashboard

AI Pipeline

Automatic transcription and metadata extraction

Search & RAG

Retrieve results via API

Post-Processing

Use summaries, chapters, and alerts

AI Features

Named Entity Recognition for people, places, and organizations

Automatic keyword and topic extraction

Speaker recognition and attribution

OCR text extraction from videos and images

Full-text and semantic search across all content

RAG-based Q&A with source references

Automated AI workflows (Flows) for insights and content generation

Multi-format transcript export (JSON, SRT, DOCX, TXT)

Auto-chapters: AI-based segmentation into thematic sections

Segment-based recommendations for contextually related content

API Integration

AI Analysis via API

Full programmatic access to all AI analysis results -- from metadata to automated workflows.

Metadata Extraction

Retrieve entities, keywords, speakers, OCR, and document content as structured data.

Full-Text & Semantic Search

Keyword-based and natural language search across your entire media library.

Chat & RAG

Natural language questions against media content -- answers with source references and timestamps.

Flows API

Parameterized AI workflows for summaries, reports, and automated content generation.

Transcript Export

Export transcripts with speaker attribution as JSON, SRT, DOCX, or TXT.

Auto-Chapters

AI segments videos automatically into thematic sections – the foundation for navigation, search, and recommendations.

Recommendation API

Recommend contextually related content at the section level – based on metadata, transcripts, and tags.

# 1. Erkannte Entitäten, Keywords und Sprecher abrufen
curl "https://api.streamdiver.com/v2/media/{assetId}/metadata?expand=entities,keywords,speakers" \\
  -H "Authorization: Bearer YOUR_TOKEN"

# 2. Semantische Suche über alle Medien
curl "https://api.streamdiver.com/v2/search/semantic?text=Welche%20Entscheidungen%20wurden%20getroffen" \\
  -H "Authorization: Bearer YOUR_TOKEN"

# 3. KI-gestützte Frage (RAG)
curl -X POST https://api.streamdiver.com/v2/chats \\
  -H "Authorization: Bearer YOUR_TOKEN" \\
  -H "Content-Type: application/json" \\
  -d '{"prompt": "Fasse die wichtigsten Punkte zusammen", "mediaIds": ["{assetId}"]}'

# 4. Flow ausführen (z.B. Meeting-Zusammenfassung)
curl -X POST https://api.streamdiver.com/v2/flows/{flowId}/runs \\
  -H "Authorization: Bearer YOUR_TOKEN" \\
  -H "Content-Type: application/json" \\
  -H "Accept: text/event-stream" \\
  -d '{"mediaId": "{assetId}", "llmProvider": "streamdiver/nitrox", "params": {"language": "de"}}'

Integration flow

Upload mediaS3 API

AI analysisWebhook

Retrieve metadataREST API

Search & ChatSemantic + RAG

Ready to get started?

Schedule Demo Open AI Docs