Live Environment Streams

A curated, agent-optimized repository of live outdoor webcam streams across the globe. Originally created to power twitch.tv/alwaysrainsomewhere.

5,172 streams · 80 countries · 19 sources · 3,027 directly usable

Data quality

Not all streams are equal. The schema includes fields that tell you exactly what you're getting:

Field	What it tells you
url_type	Whether the URL is a direct HLS manifest, a YouTube link, or an HTML page requiring browser extraction
coordinates_quality	Whether coordinates point to the actual camera (exact), a nearby city (city), or a country capital fallback (country_centroid)
status	Whether the stream has been verified working (active), not yet tested (unverified), or confirmed dead (deprecated)
source_url_requires	Special handling needed: browser for HTML pages, token_refresh for expiring auth tokens, or null for direct access

Recommended query for agent pipelines — get only streams you can use immediately:

usable = [
    f for f in streams
    if f["properties"]["url_type"] in ("hls", "youtube")
    and f["properties"]["coordinates_quality"] != "country_centroid"
    and f["properties"]["source_url_requires"] is None
]
# Returns ~2,401 streams with direct URLs and accurate coordinates

Usage

import json
import random

with open("streams.geojson", "r") as f:
    data = json.load(f)

streams = data["features"]
selected = random.choice(streams)
props = selected["properties"]
coords = selected["geometry"]["coordinates"]

print(f"Loading {props['name']} at {coords}")
print(f"URL: {props['url']} (type: {props['url_type']})")
print(f"Environment: {props['environment']}")
print(f"Source: {props['source_family']} | Coords: {props['coordinates_quality']}")

Filter by environment or quality:

# Coastal streams with direct HLS and real coordinates
coastal_hls = [
    f for f in streams
    if f["properties"]["environment"] == "coastal"
    and f["properties"]["url_type"] == "hls"
    and f["properties"]["coordinates_quality"] == "exact"
]

# All streams for a specific country
us_streams = [f for f in streams if f["properties"]["country_code"] == "US"]

GeoJSON schema

Each Feature in streams.geojson:

{
  "type": "Feature",
  "geometry": {
    "type": "Point",
    "coordinates": [-77.4875, 38.7749]
  },
  "properties": {
    "name": "I-95 at Route 123",
    "display_name": "Woodbridge",
    "url": "https://example.com/stream.m3u8",
    "country_code": "US",
    "environment": "traffic",
    "source_family": "vdot",
    "quality_tier": "standard",
    "scene_type": "traffic",
    "url_type": "hls",
    "coordinates_quality": "exact",
    "status": "unverified",
    "last_verified": null,
    "source_url_requires": null
  }
}

Property reference

Property	Type	Description
name	string	Descriptive camera name
display_name	string	Human-readable location label (city/landmark)
url	string	Stream URL (HLS manifest, YouTube link, or HTML page)
country_code	string	ISO 3166-1 alpha-2 country code
environment	string	Scene category (see table below)
source_family	string	Data source identifier (see sources.json)
quality_tier	string	standard, unknown, or low
scene_type	string	Granular scene label (traffic, beach, marina, etc.)
url_type	string	hls, youtube, html_page, or http_image
coordinates_quality	string	exact, city, or country_centroid
status	string	active, unverified, or deprecated
last_verified	string/null	ISO 8601 timestamp of last successful probe
source_url_requires	string/null	browser, token_refresh, or null

Environment values

Environment	Count	Description
traffic	2,369	Road/highway cameras
urban	1,920	City and town views
coastal	319	Beach and coastal
mountain	168	Mountain and alpine
marina	128	Marinas and harbors
waterway	103	Rivers and canals
lake	86	Lake views
nature	43	Nature and wildlife
other	36	Webcam, harbor, transport, landmark

URL types

Type	Count	How to use
hls	3,003	Pass directly to ffmpeg or yt-dlp
html_page	2,106	Requires browser/JS to extract stream URL
http_image	39	HTTP endpoint serving JPEG snapshots
youtube	24	Use yt-dlp to extract HLS manifest

Coordinates quality

Quality	Count	Accuracy
exact	3,812	Camera-level geocoded
city	303	City/interchange level (~10km)
country_centroid	1,057	Fallback to capital — do not use for geo-queries

Source breakdown

Source	Streams	Type	Countries
skylinewebcams	2,771	Scenic/tourism	65+
vdot	1,168	Virginia DOT traffic	US
mdsha	404	Maryland SHA traffic	US
autobahn_nrw	281	German Autobahn (NRW)	DE
deldot	261	Delaware DOT traffic	US
nysdot	91	New York State DOT	US
opencctv	30	Open CCTV feeds	Multi
ladot	30	Los Angeles DOT	US
youtube	24	YouTube Live	Multi
vegvesen	22	Norwegian roads	NO
ktict	20	Korea traffic	KR
iticfoundation	20	Thailand cameras	TH
taiwan_freeway	18	Taiwan highways	TW
paspro	12	Indonesia traffic	ID
vdotcameras	8	VDOT secondary	US
earthcam	6	EarthCam network	Multi
panama_canal	4	Panama Canal views	PA
quebec511	1	Quebec traffic	CA
brownrice	1	Japan	JP

Full per-source metadata in sources.json.

Country coverage (top 20)

Country	Streams	Country	Streams
US	2,255	GB	43
IT	1,055	FR	38
DE	334	HR	36
ES	267	MX	36
GR	237	CZ	31
NO	91	PE	24
BR	78	TW	24
CA	63	CH	23
MT	60	PH	23
TH	53	RO	23

80 countries total.

Directory structure

streams.geojson          # Master file (all 5,172 streams)
sources.json             # Per-source metadata (license, notes, refresh method)
data/                    # Per-country GeoJSON subsets (e.g. data/US.json)
scripts/
  validate.py            # Probe stream URLs, update status/last_verified
  geocode_swc.py         # Fix fallback coordinates for SWC streams
  regenerate_country_files.py  # Rebuild data/ from streams.geojson
  stats.py               # Print dataset statistics
.github/workflows/
  validate.yml           # Weekly automated validation

Known limitations

1. Skyline Webcams HTML pages (2,106 streams): These are HTML page URLs, not direct stream URLs. An agent needs a browser or scraper to extract the actual HLS manifest. Marked with url_type: "html_page" and source_url_requires: "browser".

2. SWC token expiry (665 streams): The 665 SWC streams with direct .m3u8 URLs use auth tokens that expire. Marked with source_url_requires: "token_refresh".

3. Fallback coordinates (1,057 streams): These streams have coordinates pointing to country capitals (Rome, NYC, Madrid, etc.) instead of actual camera locations. Marked with coordinates_quality: "country_centroid". Run scripts/geocode_swc.py to fix.

4. No live verification yet: All streams launch as status: "unverified". The GitHub Actions workflow will progressively verify them.

Contributing

To add a new source:

1. Add streams to streams.geojson following the schema above — all new fields are required.
2. Add source metadata to sources.json.
3. Run python scripts/regenerate_country_files.py to update data/.
4. Run python scripts/stats.py to verify counts.
5. Update this README with new totals.

#dataset #computer-vision #live-weather #machine-learning #webcams #hls-streams #geojson #environment-monitoring #traffic-cameras #ai-agents #geospatial