earthTV offers AI developers and machine learning teams access to a vast archive of rights-cleared, high-definition video content, optimized for AI training. Our dataset features over 100,000 minutes of footage, enriched with metadata like geolocation, timestamp, weather, sun cycle, and camera motion, covering major global cities. This makes it ideal for training models in computer vision, multimodal AI, video-language modeling, environmental AI, autonomous systems, and realistic generative AI, enabling more accurate and controllable video generation.
Rights-Cleared HD Video & Structured Metadata
At earthTV, we offer direct access to one of the largest curated archives of real-world, high-definition video content — purpose-built for AI developers and machine learning research teams.
Our dataset includes over 100,000 minutes of high-definition footage, enriched with detailed metadata: geolocation, timestamp, weather conditions, sun cycle data, and camera motion, coverage includes major global cities like New York, Berlin, Tokyo, and more..
This makes it ideal for large-scale model training in computer vision, multimodal AI, video-language modeling, environmental AI, and autonomous systems.
It is also a powerful dataset for training generative AI models that synthesize realistic video: imagine generating a sunrise over Tokyo, a rainy afternoon in Berlin, or a windy skyline view of New York — all grounded in real-world visual and environmental patterns.
By learning from authentic footage paired with rich metadata, your models can gain spatial, temporal, and weather-related context — enabling more accurate and controllable video generation.
What We Offer and Extensive Global Archive
🟣 100,000+ minutes of professionally captured HD video content.
🟣 Global coverage — diverse cities, landscapes, climates, and weather conditions.
🟣 Short, standardized clips (13–30 seconds) — perfect for batch processing and AI pipelines.
🟣 Consistent broadcast-grade quality — no black frames, stable motion, clean captures.
For Generative AI Developers
🟣 Real-world video data with full geotemporal and environmental context — perfect for training diffusion, transformer, or video synthesis models.
🟣 Enable controllable generation of scenes like “sunset in Sydney in springtime” or “snowfall in Geneva at night.”
🟣 Ground your generative models in authentic footage and structured metadata — for outputs that feel real, not imagined.
Rich, Structured Metadata
Each clip comes with a full metadata set, in machine-readable formats (JSON, CSV):
🟣 Geolocation: City, Country, Latitude, Longitude
🟣 Timestamp: Exact capture time + timezone
🟣 Weather: Temperature, humidity, wind speed/direction, cloud conditions
🟣 Sun Cycle: Sun azimuth, elevation, shadow length, daylight status, season
🟣 Camera Motion: Motion type (constant, pan), shot framing
🟣 Video Specs: HD 1920×1080, 25 fps, H.264 codec, 10 Mbps bitrate
Delivered as JSON or CSV, optimized for direct AI integration.
Access Methods
🟣 API Access: Streamline metadata and video link delivery.
🟣 Custom Delivery: Bulk datasets via Amazon S3, secure file transfer, or customized infrastructure to fit your ML pipeline.
Primary Dataset — Cleanfeed Clips
✅ 13–30 second clips, captured continuously from fixed locations worldwide.
✅ Full metadata attached for supervised learning, temporal labeling, and multimodal AI models.
✅ Standardized, high-quality format ensures training efficiency at scale.
✅ Perfect for large-scale vision models, geo-temporal analysis, and environmental AI projects.
▶️ Preview Cleanfeed Clip:
📥 View Cleanfeed Clip + Metadata (JSON)
{
"type": "clip",
"source": {
"id": "GX6kbzOAB-0",
"alias": "WAS",
"createTime": "2019-01-09T09:18:03.333207Z",
"updateTime": "2025-05-07T08:05:50.247479Z",
"accountId": "earthtv",
"title": {
"en": "Washington D.C. Capitol"
},
"description": {
"en": ""
},
"location": {
"timeZone": "America/New_York",
"coordinates": {
"lat": 38.8969,
"lng": -77.0093
},
"elevation": 12.7109,
"city": {
"en": "Washington D.C.",
},
"cityId": 210,
"state": {
"en": "District of Columbia",
},
"stateId": 209,
"country": {
"en": "USA",
},
"countryId": 8,
"continent": {
"en": "North America",
},
"continentId": 4
},
"tags": [],
"userData": {
},
"region": "us",
},
"medium": {
"@type": "earthtv.com/earthtv.loe.media.Clip",
"id": "MxITfO0ABKc",
"alias": "WAS_20250605_172012.mp4",
"createTime": "2025-06-05T21:22:12.533612Z",
"updateTime": "2025-06-05T21:23:09.915455Z",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"tags": [],
"userData": {
},
"region": "us",
"duration": 15080,
"startTime": "2025-06-05T17:20:13-04:00",
"timeZone": "America/New_York",
"ttl": 0,
"sourceVariantId": "MqItfu48bTt",
"status": "finished",
"quality": {
"blackFrames": "false"
},
"maintenance": false,
"type": "footage",
"mediaUrls": {
"previewImage": {
}
},
"storage": ""
},
"moments": [
{
"id": "MqI3f68AjKk",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"time": "2025-06-05T17:20:13-04:00",
"timeZone": "America/New_York",
"provider": "system",
"clipId": "MxITfO0ABKc",
"sun": {
"azimuth": 272.77597,
"elevation": 34.298695,
"declination": 22.643347,
"conditions": [
"day"
],
"shadowLength": 1.4660169,
"season": "summer"
},
"custom": {}
},
{
"id": "MqItgQKABK8",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"timeZone": "UTC",
"provider": "camera",
"clipId": "MxITfO0ABKc",
"custom": {},
"trajectory": {
"name": "WAS_15_12",
"type": "constant_motion",
"shot": "full",
"movement": [
"down_up"
]
}
},
{
"id": "MqItgIaABK4",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"timeZone": "UTC",
"provider": "camera",
"clipId": "MxITfO0ABKc",
"weather": {
"observationTime": "2025-06-05T21:15:30Z",
"temperature": 28.9,
"windSpeed": 7.1,
"windDirection": 173,
"condition": "scattered_02",
"humidity": 65.3
},
"custom": {}
},
{
"id": "MqIt7z-ABPA",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"time": "2025-06-05T21:23:09.986984Z",
"timeZone": "UTC",
"provider": "system",
"clipId": "MxITfO0ABKc",
"custom": {},
"quality": {
"blackFrames": false,
"broadcast": true,
"visual": "unknown"
}
}
],
"variants": [
{
"id": "MqItfu48bTt",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"clipId": "MxITfO0ABKc",
"createTime": "2025-06-05T21:22:12.53511Z",
"updateTime": "2025-06-05T21:22:12.53511Z",
"container": "mp4",
"videoCodec": "h264",
"videoProfile": "high@l4",
"videoWidth": 1920,
"videoHeight": 1080,
"videoFramerate": 25,
"videoBitrate": 10240000,
"videoGop": 0,
"audioBitrate": 0,
"contentType": "video/mp4",
"url": "",
"region": "us",
"status": "finished",
"size": 14334086,
"urls": {
"edge": "",
"cdn": ""
}
}
]
}{
"type": "clip",
"source": {
"id": "GX6kbzOAB-0",
"alias": "WAS",
"createTime": "2019-01-09T09:18:03.333207Z",
"updateTime": "2025-05-07T08:05:50.247479Z",
"accountId": "earthtv",
"title": {
"en": "Washington D.C. Capitol"
},
"description": {
"en": ""
},
"location": {
"timeZone": "America/New_York",
"coordinates": {
"lat": 38.8969,
"lng": -77.0093
},
"elevation": 12.7109,
"city": {
"en": "Washington D.C.",
},
"cityId": 210,
"state": {
"en": "District of Columbia",
},
"stateId": 209,
"country": {
"en": "USA",
},
"countryId": 8,
"continent": {
"en": "North America",
},
"continentId": 4
},
"tags": [],
"userData": {
},
"region": "us",
},
"medium": {
"@type": "earthtv.com/earthtv.loe.media.Clip",
"id": "MxITfO0ABKc",
"alias": "WAS_20250605_172012.mp4",
"createTime": "2025-06-05T21:22:12.533612Z",
"updateTime": "2025-06-05T21:23:09.915455Z",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"tags": [],
"userData": {
},
"region": "us",
"duration": 15080,
"startTime": "2025-06-05T17:20:13-04:00",
"timeZone": "America/New_York",
"ttl": 0,
"sourceVariantId": "MqItfu48bTt",
"status": "finished",
"quality": {
"blackFrames": "false"
},
"maintenance": false,
"type": "footage",
"mediaUrls": {
"previewImage": {
}
},
"storage": ""
},
"moments": [
{
"id": "MqI3f68AjKk",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"time": "2025-06-05T17:20:13-04:00",
"timeZone": "America/New_York",
"provider": "system",
"clipId": "MxITfO0ABKc",
"sun": {
"azimuth": 272.77597,
"elevation": 34.298695,
"declination": 22.643347,
"conditions": [
"day"
],
"shadowLength": 1.4660169,
"season": "summer"
},
"custom": {}
},
{
"id": "MqItgQKABK8",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"timeZone": "UTC",
"provider": "camera",
"clipId": "MxITfO0ABKc",
"custom": {},
"trajectory": {
"name": "WAS_15_12",
"type": "constant_motion",
"shot": "full",
"movement": [
"down_up"
]
}
},
{
"id": "MqItgIaABK4",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"timeZone": "UTC",
"provider": "camera",
"clipId": "MxITfO0ABKc",
"weather": {
"observationTime": "2025-06-05T21:15:30Z",
"temperature": 28.9,
"windSpeed": 7.1,
"windDirection": 173,
"condition": "scattered_02",
"humidity": 65.3
},
"custom": {}
},
{
"id": "MqIt7z-ABPA",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"time": "2025-06-05T21:23:09.986984Z",
"timeZone": "UTC",
"provider": "system",
"clipId": "MxITfO0ABKc",
"custom": {},
"quality": {
"blackFrames": false,
"broadcast": true,
"visual": "unknown"
}
}
],
"variants": [
{
"id": "MqItfu48bTt",
"accountId": "earthtv",
"sourceId": "GX6kbzOAB-0",
"clipId": "MxITfO0ABKc",
"createTime": "2025-06-05T21:22:12.53511Z",
"updateTime": "2025-06-05T21:22:12.53511Z",
"container": "mp4",
"videoCodec": "h264",
"videoProfile": "high@l4",
"videoWidth": 1920,
"videoHeight": 1080,
"videoFramerate": 25,
"videoBitrate": 10240000,
"videoGop": 0,
"audioBitrate": 0,
"contentType": "video/mp4",
"url": "",
"region": "us",
"status": "finished",
"size": 14334086,
"urls": {
"edge": "",
"cdn": ""
}
}
]
}Extended Context Sample — Program Clips
✅ 90+ seconds of footage for projects requiring longer video sequences.
✅ Supports research in video summarization, action recognition, video-to-text
▶️ Preview Program Clips
Sample Metadata Preview
| Field | Example |
|---|---|
| Title | Washington D.C. Capitol |
| Latitude | 38.8969 |
| Longitude | -77.0093 |
| Date | 2025-06-05 |
| Time | 17:20:13 (America/New_York) |
| Temperature | 28.9°C |
| Humidity | 65.3% |
| Wind Speed | 7.1 km/h |
| Weather Condition | Scattered Clouds |
| Sun Elevation | 34° (Daylight, Summer) |
| Movement Type | Constant Motion (Full Shot, Down-Up) |
| Video Format | MP4, H.264, 1920×1080, 25 fps |
| Bitrate | 10.24 Mbps |
| Duration | 15 seconds |
| Quality | Broadcast-grade, No Black Frames |
Why earthTV?
✅ Original Rights Holder — No intermediaries.
✅ Fully Rights-Cleared — Ready for commercial AI model training and research.
✅ Rich Real-World Data — From urban to natural landscapes, real weather, true sun conditions.
✅ Scalable Licensing Options — Non-exclusive, tailored to your dataset size and project needs.
✅ Seamless Access — API or secure bulk transfer for large-scale ingestion.