Nearly every panel on the TerminalFeed dashboard runs on a free, public endpoint. Bitcoin ticks, earthquake alerts, the ISS position, Steam player counts: almost none of it costs money, as long as you respect the provider's limits. Here is what we use, with endpoints, current limits, key requirements and our honest take.
We re-tested every entry in September 2026, and plenty moved. Reddit closed its keyless .json door in late May. abuse.ch has required a free key since June 2025. Open Notify's astronaut list still names a 2024 crew, and SteamSpy's player counts froze. Frankfurter changed domains and grew from about 30 currencies to 206. Where an API died or rotted, the card names its replacement.
> CRYPTO & FINANCE
wss://ws-feed.exchange.coinbase.comRate Limit: No auth for public channels; subscribe within 5 seconds or get dropped
Key Required: No
wss://stream.binance.com:9443/ws/ (data only: wss://data-stream.binance.vision)Rate Limit: 5 incoming messages/sec, 1,024 streams per connection, 300 connections per 5 min per IP; connections drop at 24 hours
Key Required: No
api.kraken.com/0/public/ + wss://ws.kraken.com/v2Rate Limit: About 150 WebSocket connection attempts per 10 min per IP, then a 10-minute ban
Key Required: No
api.coingecko.com/api/v3/Rate Limit: Keyless calls share a per-IP throttle; the free Demo key gets 100 calls/min and 10,000/month
Key Required: Free Demo key recommended (attribution required)
api.coinlore.net/api/tickers/Rate Limit: No hard limit; about 1 request/sec recommended
Key Required: No
POST api.hyperliquid.xyz/infoRate Limit: 1,200 weight/min per IP;
allMids costs 2, most info requests 20Key Required: No
api.alternative.me/fng/Rate Limit: None documented
Key Required: No
mempool.space/api/Rate Limit: Enforced but unpublished; 429s for excess, bans for repeat offenders
Key Required: No
api.frankfurter.dev/v2/ratesRate Limit: No quotas; abuse is throttled
Key Required: No
api.frankfurter.app now 301-redirects to the deprecated but still running v1.> NEWS & SOCIAL
hacker-news.firebaseio.com/v0/Rate Limit: "There is currently no rate limit," per the docs
Key Required: No
public.api.bsky.app/xrpc/Rate Limit: Cached and "generous," no published number (3,000 per 5 min per IP applies to authenticated PDS traffic)
Key Required: No (this host does not accept auth)
reddit.com/r/{subreddit}/.rssRate Limit: Tight; watch the
x-ratelimit-remaining headerKey Required: No (
.json now returns 403 without OAuth)
.json in late May 2026. RSS still worked in September. Don't build anything critical on it.> WORLD DATA
earthquake.usgs.gov/earthquakes/feed/v1.0/summary/Rate Limit: None; static files regenerated every minute
Key Required: No
api.open-meteo.com/v1/forecastRate Limit: 600/min, 5,000/hr, 10,000/day, 300,000/month
Key Required: No
api.weather.govRate Limit: Unpublished; throttled calls can usually retry after about 5 seconds
Key Required: No, but a User-Agent naming your app is mandatory
/alerts/active is the best free severe-weather feed in the States.gdacs.org/xml/rss.xmlRate Limit: None documented
Key Required: No
wss://www.seismicportal.eu/standing_order/websocketRate Limit: None documented
Key Required: No
> DEV & TECH
api.github.comRate Limit: 60 req/hr unauthenticated, 5,000 with a token; unauthenticated search is 10/min
Key Required: No, but use a token
api.stackexchange.com/2.3/Rate Limit: 300 req/day keyless, 10,000 with a free key; over 30 req/sec per IP gets banned
Key Required: No (a free key helps a lot)
backoff field. Keyless quota is shared by every app on your IP.{status-domain}/api/v2/status.jsonRate Limit: None documented
Key Required: No
summary.json for per-component detail. It is how our status monitor works.> SPACE
api.wheretheiss.at/v1/satellites/25544 + corquaid.github.io/international-space-station-APIs/JSON/people-in-space.jsonRate Limit: Roughly 1 req/sec for the position API
Key Required: No
ll.thespacedevs.com/2.3.0/Rate Limit: 15 req/hr free (check yours at
/2.3.0/api-throttle/)Key Required: No
api.nasa.gov/planetary/apodRate Limit: 1,000 req/hr with your own key; DEMO_KEY allows 30/hr and 50/day per IP
Key Required: Yes (free, instant)
services.swpc.noaa.gov/json/Rate Limit: None documented
Key Required: No
> SECURITY
urlhaus-api.abuse.ch/v1/ + threatfox-api.abuse.ch/api/v1/Rate Limit: Fair use; bulk dumps refresh every 5 minutes
Key Required: Yes, a free Auth-Key (mandatory since June 30, 2025)
cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.jsonRate Limit: None (static file, CSV too)
Key Required: No
internetdb.shodan.io/{ip}Rate Limit: None published; data refreshes weekly
Key Required: No
> PREDICTION MARKETS
gamma-api.polymarket.comRate Limit: 4,000 req/10s overall, 300 req/10s on
/markets; excess is queued, not rejectedKey Required: No
> ENTERTAINMENT
api.steampowered.com/ISteamChartsService/GetGamesByConcurrentPlayers/v1/Rate Limit: None published
Key Required: No
dog.ceo/api/breeds/image/random + api.thecatapi.com/v1/images/searchRate Limit: Dog CEO documents none; the Cat API's free plan is 10,000 requests/month
Key Required: No (a free Cat API key for real volume)
> REFERENCE
en.wikipedia.org/api/rest_v1/feed/onthisday/ + stream.wikimedia.org/v2/stream/recentchangeRate Limit: New in 2026: 10 req/min unidentified, 200 req/min with a proper User-Agent
Key Required: No
zenquotes.io/api/randomRate Limit: 5 req per 30 sec per IP
Key Required: No (attribution link required)
/api/quotes batch once and rotate locally instead of calling per page view.> How We Keep 30+ Upstreams Honest
Calling this many APIs straight from a browser would burn through half these limits before lunch. The dashboard never does. Every request goes through our own backend, which caches each upstream on its own schedule, shares one answer with every visitor, and serves the last good copy when a provider throttles or goes dark. The strictest feeds are fetched on a timer, so a traffic spike never becomes a burst of upstream calls.
We also watch for the quieter failure: a feed that still answers 200 but has stopped changing. SteamSpy and Open Notify are textbook cases, and the FEED HEALTH panel on the dashboard tracks it live.
Building your own dashboard, bot or agent? Cache first, attribute where asked, and keep a second source for anything you can't live without. Our rate limit guide covers the mechanics. Happy building.