ESPN Scraper

Scrape ESPN scores, schedules, odds, standings, game summaries, team rosters, rankings, and news from public sports pages. Export ESPN data to JSON, CSV, or a spreadsheet with one record per game, team, player, article, or poll entry. This is an ESPN data API alternative for public information across several major sports and soccer leagues.

The actor uses structured public data endpoints. It does not need a login or a browser session. Scoreboard requests are split by date, so a date range works even though the underlying endpoint accepts one day per request. Every record identifies its type and league, making it possible to keep separate tables or join data by game, team, and player IDs.

What data can you extract from ESPN?

The selected mode controls the dataset rows. Scoreboard mode produces games and scores. Game mode produces one detailed summary per requested event. Odds mode emits a game with a flattened published line. Standings, Teams, Roster, Schedule, News, and Rankings modes each emit their corresponding record type. Fields tied to a particular sport or game state may be absent or null. For example, a scheduled game has no final score and many completed games have no current betting line.

Field	Description	Example
type	Record kind: game, gameSummary, oddsLine, standing, team, player, scheduleGame, article, or ranking	game
league	Requested league key	college-football
leagueName	Full league name	NCAA - Football
gameId	ESPN event ID	401856704
date	Game start time	2026-09-26T16:00Z
name	Event display name	Texas Longhorns at Tennessee Volunteers
shortName	Short event name	TEX @ TENN
status	State, detail, completion, period, and clock	{"state":"in","period":2}
homeTeam	Home team name	Tennessee Volunteers
homeTeamId	Home team ID	2633
homeAbbr	Home team abbreviation	TENN
homeScore	Home score	3
homeLogo	Home logo URL	https://a.espncdn.com/...
homeRecord	Home team record	3-0
awayTeam	Away team name	Texas Longhorns
awayTeamId	Away team ID	251
awayAbbr	Away team abbreviation	TEX
awayScore	Away score	10
awayLogo	Away logo URL	https://a.espncdn.com/...
awayRecord	Away team record	3-0
odds	Primary sportsbook odds or full summary lines	{"spread":-7.0,...}
venue	Venue name, location, and indoor flag	{"name":"Neyland Stadium",...}
broadcasts	Broadcast network names	["TV network"]
leaders	Published game leaders	[...]
tickets	Public ticket summary	null
situation	Live game situation	{"down":2,...}
competitors	Competitors and published statistics	[...]
url	ESPN page URL	https://www.espn.com/...
scrapedAt	UTC collection time	2026-09-26T17:05:51Z
gameDetails	Optional game summary attached to a scoreboard row	{"type":"gameSummary",...}
provider	Sportsbook name	sportsbook.example.com
details	Line summary	home -7
spread	Current home point spread	-7.0
overUnder	Current total points line	50.5
homeMoneyLine	Current home money line	-345
awayMoneyLine	Current away money line	275
openSpread	Opening home point spread	-2.5
openOverUnder	Opening total points line	48.5
openHomeMoneyLine	Opening home money line	-142
openAwayMoneyLine	Opening away money line	120
spreadMovement	Current spread minus opening spread	-4.5
totalMovement	Current total minus opening total	2.0
homeSpreadPrice	Current home spread price	-115
awaySpreadPrice	Current away spread price	-105
overPrice	Current over price	-102
underPrice	Current under price	-118
betUrl	Betting line link	https://sportsbook.example.com/...
teamStats	Team box scores and statistics	[...]
playerStats	Player box scores and statistics	[...]
winProbability	Published win probability feed	[...]
seasonSeries	Season head-to-head information	[...]
injuries	Published injuries	[...]
officials	Game officials	[...]
attendance	Game attendance	18997
plays	Optional play-by-play records	[...]
season	Season year	2026
group	Conference or division	Eastern Conference
teamId	Team ID	13
team	Team name	Los Angeles Lakers
abbreviation	Team abbreviation	LAL
logo	Team logo URL	https://a.espncdn.com/...
stats	Named standings statistics	{"wins":"41",...}
wins	Wins	41
losses	Losses	29
rank	Standing or poll rank	1
playerId	Player ID	5113969
position	Player position	SG
jersey	Jersey number	7
age	Player age	21
height	Height in inches	77
weight	Weight in pounds	184
college	College name where published	null
salary	Salary where published	null
location	Team location	Los Angeles
slug	Team URL slug	los-angeles-lakers
color	Team color	552583
record	Team record	41-29
standingSummary	Team standing summary	Pacific Division
nextEvent	Next scheduled event	[...]
forTeamId	Team whose schedule was requested	13
forTeamResult	W when that team won a completed game	W
articleId	Article ID	50034054
headline	Article headline	Public sports headline
description	Article summary	Article summary
publishedAt	Article publication time	2026-09-26T12:00Z
updatedAt	Article update time	2026-09-26T13:00Z
categories	Article topic tags	[...]
poll	Poll name	AP Top 25
pollDate	Poll date	2026-09-20T07:00Z
week	Poll week	Week 4
previousRank	Previous poll rank	2
points	Poll points	1706
providers	All available provider lines returned for this game	[{"spread":-7.0,...}]

The game fields are deliberately stable across sports. A soccer event may contain individual competitors or extra status details in competitors, while a basketball game commonly has team scores and period information. The original nested arrays are kept when they carry useful sport-specific detail. status, venue, and odds are compact nested objects; they remain objects when exported as JSON and may be serialized as text in flat formats.

How to scrape ESPN

1. Pick a mode. Start with Scoreboard if you want the games on a date, or Standings if you want a current league table.
2. Enter one or more supported league keys. For a single game, use eventIds. For a roster or team schedule, use teamIds. You can also paste supported page URLs into startUrls.
3. Set dates for Scoreboard or Odds. Use a single calendar day, a relative day, or a range up to 32 days inclusive. Set maxResults to cap the number of records charged and returned.
4. Start the run. When it finishes, download the dataset in JSON, CSV, or spreadsheet format. Repeat with a different mode if you need a different table.

A typical workflow collects a scoreboard first, reads its gameId values, and then requests Game mode for selected events. Another workflow fetches a team directory, reads teamId, and then requests rosters or schedules. The ID fields preserve these joins without comparing names, which can vary by league or season.

Input

Parameter	Type	Default	Description
mode	string	scoreboard	Scoreboard, game, odds, standings, teams, roster, schedule, news, or rankings.
leagues	string array	nba in the form	League keys for league-wide modes.
dates	string	Today if blank	YYYYMMDD, YYYY-MM-DD, today, yesterday, tomorrow, or YYYYMMDD-YYYYMMDD.
eventIds	string array	Empty	Game IDs prefixed by league, such as nfl/401872953.
teamIds	string array	Empty	Team numeric IDs or abbreviations prefixed by league, such as nba/13 or nba/LAL.
startUrls	string array	Empty	ESPN game, standings, scoreboard, or team URLs.
enrichGameDetails	boolean	false	Fetch and attach a summary for each scoreboard game.
includePlayByPlay	boolean	false	Include play records in Game mode or attached game details.
enrichTeamDetails	boolean	false	Fetch full team profiles instead of only the directory entries.
season	string	Current source season	Year for standings, schedules, or rankings.
seasonType	string	Source default	Schedule season type: 1 preseason, 2 regular season, 3 postseason.
groups	string	None	Source group ID for college scoreboard filtering.
maxResults	integer	200	Overall record cap.
maxConcurrency	integer	5	Maximum concurrent requests, capped at 10.
proxyConfiguration	object	Enabled	Proxy configuration for endpoint requests.

{
  "mode": "scoreboard",
  "leagues": ["college-football"],
  "dates": "20260926",
  "groups": "80",
  "maxResults": 65
}

Supported league keys include nfl, college-football, ncaaf, ufl, nba, wnba, ncaam, ncaaw, mlb, nhl, epl, la-liga, serie-a, bundesliga, ligue-1, mls, ucl, uel, liga-mx, world-cup, nwsl, f1, ufc, and pga. These are key mappings, not a guarantee that every mode has records for every league. Polls, for instance, are published in selected sports. Empty dates during an off-season normally produce no scoreboard rows.

Output example

This trimmed item came from a live scoreboard run for 26 September 2026. The run returned 65 games with game ID, date, teams, and URL filled on every row. Live scores and the clock can change between runs.

{
  "type": "game",
  "league": "college-football",
  "leagueName": "NCAA - Football",
  "gameId": "401856704",
  "date": "2026-09-26T16:00Z",
  "name": "Texas Longhorns at Tennessee Volunteers",
  "status": {"state": "in", "detail": "12:07 - 2nd Quarter", "completed": false, "period": 2, "clock": "12:07"},
  "homeTeam": "Tennessee Volunteers",
  "homeTeamId": "2633",
  "homeScore": 3,
  "awayTeam": "Texas Longhorns",
  "awayTeamId": "251",
  "awayScore": 10,
  "venue": {"name": "Neyland Stadium", "city": "Knoxville", "state": "TN", "id": "3853", "indoor": false},
  "odds": null,
  "url": "https://www.espn.com/college-football/game?gameId=401856704"
}

Use cases

• A sports results archive can collect completed games by date and join details through gameId.
• A live display can refresh one league's scoreboard and show status, clock, teams, and score.
• A standings report can track wins, losses, and the source's full named statistics for each club.
• A roster workflow can combine player IDs, positions, measurements, and public team identifiers.
• A schedule calendar can use event dates and venue details, then mark completed results.
• A betting analysis table can compare a published current spread or total with its opening figure where both exist.

How much does it cost to scrape ESPN?

The current event price is $0.001 per dataset result, or $1 per 1,000 results. A run with 100 output rows costs $0.10 in result charges; 1,000 rows cost $1; 5,000 rows cost $5. Set maxResults to keep the maximum number of result charges predictable. A run can also consume platform resources such as compute and proxy traffic, depending on the account and configuration. The current run price shown before starting a task is authoritative if pricing changes.

A scoreboard game, a standing row, and a player row each count as one result when written to the dataset. Fetching a game summary to attach to a scoreboard row adds a request but does not create an additional result by itself. Odds mode only emits games that have a published line, so a league can produce fewer paid results than its total number of scheduled games.

Tips for best results

Use a date that falls in the league's season. today is convenient for recurring tasks, while an explicit date gives reproducible historical queries. When comparing live scores, save successive datasets with their scrapedAt value; a new run can contain changed scores. This actor does not maintain cross-run memory or output only changes.

For large historical windows, run several ranges of at most 32 days. The actor expands each range into daily scoreboard calls and deduplicates repeated game IDs within a run. maxResults stops output across all requested leagues and dates. Put the most important leagues first when the limit is smaller than the expected combined volume.

Use numeric team IDs when possible. Abbreviations are resolved through the league team directory, which adds a request. A full team profile also adds one request for each selected team. enrichGameDetails adds a summary request for each scoreboard game; use it only when the box score, injuries, or probability feed is needed. Play-by-play records can be large, so leave includePlayByPlay off for compact exports.

For college football, groups can narrow the scoreboard to a subdivision. The golden case used group 80 and produced 65 games for its date. Other group IDs may yield different records. If a league unexpectedly returns no games, check the season, calendar date, and group filter before increasing limits.

Limits

The source's scoreboard accepts a single day per request. Its response did not honor a very high limit consistently: a request with limit=1000 yielded 25 games for the golden date, while limit=200 yielded 65. The actor uses 200 and splits date ranges into daily calls. Ranges longer than 32 days are rejected so a typo does not cause a large request burst.

Odds are context dependent. The scheduled NFL sample returned lines for 14 games; a live college football sample had no primary odds in its scoreboard response. The tested game summary supplied one provider line. providers therefore contains the available line or lines; it does not promise every sportsbook on the market. When the source omits a provider label but includes a betting link, provider contains the link host. Opening lines and movement remain null when their source values are absent.

Game summaries for future events can have empty player statistics and no final score. A standalone summary may also omit an exact ISO game date even though its scoreboard counterpart has one. Team rosters and article feeds depend on what the league publishes. Rankings are only meaningful where polls exist; a ranking team name may require a directory lookup. The actor does not scrape restricted articles or require a login. Only public metadata is included for news.

Automated workflows

A client can request “Get tomorrow's NFL games with scores and published odds” by using Scoreboard mode, nfl, and tomorrow. Another can request “Get the current college football standings and a team roster” as two tasks, one for Standings and one for Roster. The resulting JSON has stable IDs and a type value, so the records can be routed into separate tables without inspecting every nested field.

FAQ

Is this an official ESPN service? No. It collects public data and has no affiliation with the source. Use the data in line with the source's terms and applicable law, including rules on personal data and downstream redistribution.

Do I need proxies or an account on the target site? The observed public JSON endpoints worked without a target-site login or browser challenge. Proxy configuration is available if an endpoint blocks a particular run. A target account is not required for the supported public modes.

How fast is a run? A scoreboard request for a single day needs one source call per league. Detail enrichment makes one extra call per game, and a date range makes one scoreboard call per day and league. Actual duration also includes startup, network response time, and output writes.

Can I schedule it or monitor changes? You can run the same input on a schedule and compare gameId, scores, status, odds, and scrapedAt across datasets. The actor does not store a cross-run change ledger or suppress unchanged records.

Can I export to a spreadsheet? Yes. Download the dataset in JSON or CSV and open the CSV in a spreadsheet. Nested fields such as odds and stats are easiest to inspect in JSON.

Why is a field empty? The site may not publish it for that game state, sport, league, or season. A future game has no final score; an event without a betting market has no odds. Some roster fields, including college and salary, are not universal.

Why are there fewer Odds rows than Scoreboard games? Odds mode deliberately skips events without a line. Scoreboard mode keeps those games with odds: null, preserving the full schedule and scores.

Integrations

Use the dataset endpoint for downstream pipelines or export the results after each run. Scheduled runs can feed webhooks, spreadsheet imports, reporting jobs, and other automation systems. The stable league, gameId, teamId, playerId, and articleId fields help keep updates aligned across recurring extracts.

Support

Open an issue on the actor's Issues tab with the mode, league, date, and a sample missing field. That context makes it possible to reproduce a source change or add a useful field.