Find any moment instantly
Find any moment instantly
Search across hours of video with natural language.
Query scenes, speech, and custom indexes in real-time.
Search across hours of video with natural language. Query scenes, speech, and custom indexes in real-time.
Timeline
6 moments found
00:12
94%
02:15
91%
03:42
89%
00:12
94%
03:54
67%
00:36
94%
Timeline
6 moments found
00:12
94%
02:15
91%
03:42
89%
00:12
94%
03:54
67%
00:36
94%
Timeline
6 moments found
00:12
94%
02:15
91%
03:42
89%
00:12
94%
03:54
67%
00:36
94%
Real time now, searchable later
Real time now, searchable later
Real time now, searchable later
For streams, you can paginate scenes as they are created. You can also keep the index for historical querying and replay as "episodic memory."
For streams, you can paginate scenes as they are created. You can also keep the index for historical querying and replay as "episodic memory."
from videodb import connect, IndexType # Connect to VideoDB conn = connect() rtstream = conn.get_rtstream("ops_monitor_v1") # Search across live stream with scene index results = rtstream.search( query="forklift near loading dock", index_type=IndexType.scene, result_threshold=0.8 ) # Process real-time results for hit in results: print(f"[{hit.timestamp}] Score: {hit.score}")
from videodb import connect, IndexType # Connect to VideoDB conn = connect() rtstream = conn.get_rtstream("ops_monitor_v1") # Search across live stream with scene index results = rtstream.search( query="forklift near loading dock", index_type=IndexType.scene, result_threshold=0.8 ) # Process real-time results for hit in results: print(f"[{hit.timestamp}] Score: {hit.score}")
from videodb import connect, IndexType # Connect to VideoDB conn = connect() rtstream = conn.get_rtstream("ops_monitor_v1") # Search across live stream with scene index results = rtstream.search( query="forklift near loading dock", index_type=IndexType.scene, result_threshold=0.8 ) # Process real-time results for hit in results: print(f"[{hit.timestamp}] Score: {hit.score}")
.py
from videodb import connect, IndexType # Connect to VideoDB conn = connect() rtstream = conn.get_rtstream("ops_monitor_v1") # Search across live stream with scene index results = rtstream.search( query="forklift near loading dock", index_type=IndexType.scene, result_threshold=0.8 ) # Process real-time results for hit in results: print(f"[{hit.timestamp}] Score: {hit.score}")
Index Stack
Index Stack
Index Stack
Layer Multiple Indexes to Power Your Search
Layer Multiple Indexes to Power Your Search
Scene Index
Scene Index
visual
Visual scene understanding and object detection
Speech Index
Speech Index
speech
Transcribed audio and spoken content
Safety Alerts
Safety Alerts
custom
Custom safety violation detection model
OCR Text
OCR Text
text
On-screen text and signage extraction
Bring your own model or metadata to index
Bring your own model or metadata to index
Multimodal Search
Multimodal Search
Multimodal Search
Combine Visual and Speech Indexes for Comprehensive Results
Combine Visual and Speech Indexes for Comprehensive Results
Python
# Multimodal search configuration results = video.search( query="safety incident", index_type=[IndexType.scene, IndexType.speech], strategy="intersection", weights={"visual": 0.6, "speech": 0.4} )
Search Modes
Search Modes
Search Modes
Choose the Right Search Strategy for Your Use Case
Choose the Right Search Strategy for Your Use Case
Natural
Semantic
Keyword
Deep Search
Semantic Search
Deep meaning understanding with context awareness
Speed
Medium
Accuracy
Very High
Best for
Complex queries requiring scene understanding
# Set search mode results = video.search(query, mode="semantic")
Natural
Semantic
Keyword
Deep Search
Semantic Search
Deep meaning understanding with context awareness
Speed
Medium
Accuracy
Very High
Best for
Complex queries requiring scene understanding
# Set search mode results = video.search(query, mode="semantic")
Natural
Semantic
Keyword
Deep Search
Semantic Search
Deep meaning understanding with context awareness
Speed
Medium
Accuracy
Very High
Best for
Complex queries requiring scene understanding
# Set search mode results = video.search(query, mode="semantic")
Natural
Semantic
Keyword
Deep Search
Semantic Search
Deep meaning understanding with context awareness
Speed
Medium
Accuracy
Very High
Best for
Complex queries requiring scene understanding
# Set search mode results = video.search(query, mode="semantic")
Search returns playable streams, not just metadata
Search returns playable streams, not just metadata
Search returns playable streams, not just metadata
Turn results into playable segments and programmatically compose them into a single stream
Go beyond timestamps and thumbnails. Get actual video clips you can play, Review, and compile into a single stream—no manual scrubbing required.
Search Results
Search Results
95%
03:42
6s
Vehicle movement detected
86%
03:42
20s
Forklift entering zone A
75%
04:18
6s
Safety zone breach
73%
05:33
6s
Equipment operation cycle
Instant Playback Stream
# Get actual video stream, not just timestamps stream = search_results.compile() stream.play() # Instantly watchable video
Stream link
https://stream.videodb.io/compiled/abc123xyz
Metadata Filters
Metadata Filters
Metadata Filters
Narrow Down Results With Structured Metadata Filters
Narrow Down Results With Structured Metadata Filters
Filters Available :
camera_view
site
time_range
tag
Active Filters
3
site: warehouse_east
camera_view: cam_dock_01
camera_view: cam_exit_01
tag: high_priority
Add Filter
1
#1Apply metadata filters
2
results = video.search (
3
query= "forklift activity",
4
filters= {
5
"site" : "warehouse_east"
6
"camera_view" : "cam_dock_01","cam_exit_01"
"camera_view" : "cam_dock_01",
7
"tag" : "high_priority"
"cam_exit_01"
8
"tag" : "high_priority"
8
9
})
Search Architecture
Search Architecture
Search Architecture
End-to-End Pipeline From Media Ingestion to Agent Actions
End-to-End Pipeline From Media Ingestion to Agent Actions
Media
Media
Media
Video streams & uploads
Video streams & uploads
Indexes
Indexes
Indexes
Scene, Speech, Custom
Scene, Speech, Custom
Search
Search
Search
Semantic query engine
Semantic query engine
Agent Loop
Agent Loop
Automated workflows
Trigger alerts on match
Auto-compile highlight reels
Feed to downstream APIs
Ready to Build Smarter Search?
Ready to Build Smarter Search?
Start Indexing Your Video Content and Unlock Semantic Search Capabilities.
Start indexing your video content and unlock semantic search capabilities
FAQs
FAQs
FAQs
What is Search in VideoDB, conceptually?
Search is the retrieval contract over time based media. You query in natural language, and you get back ranked moments with timestamps, metadata, and a playable stream window for verification.
What objects can I search?
Why is Search index driven?
Can I use more than one index in a single query?
What is Search in VideoDB, conceptually?
Search is the retrieval contract over time based media. You query in natural language, and you get back ranked moments with timestamps, metadata, and a playable stream window for verification.
What objects can I search?
Why is Search index driven?
Can I use more than one index in a single query?
What is Search in VideoDB, conceptually?
Search is the retrieval contract over time based media. You query in natural language, and you get back ranked moments with timestamps, metadata, and a playable stream window for verification.
What objects can I search?
Why is Search index driven?
Can I use more than one index in a single query?
What is Search in VideoDB, conceptually?
What objects can I search?
Why is Search index driven?
Can I use more than one index in a single query?
The Perception Layer for AI
Apt 2111 Lansing Street San Francisco, CA 94105 USA
HD-239, WeWork Prestige Atlanta, 80 Feet Main Road, Koramangala I Block, Bengaluru, Karnataka, 560034
sales@videodb.com
AUTOMATION
RESOURCES
The Perception Layer for AI
Apt 2111 Lansing Street San Francisco, CA 94105 USA
HD-239, WeWork Prestige Atlanta, 80 Feet Main Road, Koramangala I Block, Bengaluru, Karnataka, 560034
sales@videodb.com
AUTOMATION
RESOURCES
The Perception Layer for AI
Apt 2111 Lansing Street San Francisco, CA 94105 USA
HD-239, WeWork Prestige Atlanta, 80 Feet Main Road, Koramangala I Block, Bengaluru, Karnataka, 560034
sales@videodb.com