Flashscore Scraper - Live Scores, Fixtures, Stats & H2H (14 Sports)

Scrape Flashscore match data across 14 sports from one Actor: football, tennis, basketball, ice hockey, American football, baseball, handball, bandy, futsal, volleyball, cricket, darts, snooker and badminton. Pull an entire day of fixtures and results per sport, watch only what is in play right now, or open a single match and get expected goals (xG), ball possession, shot maps, goals and cards minute by minute, lineups with player ratings, and head-to-head history.

No account, no API key, no cookies, no browser. The Actor talks to the same public data feed the Flashscore website itself loads, so a run costs a handful of plain HTTP requests instead of a headless browser.

Why use this actor

• One Actor, 14 sports. Most Flashscore scrapers on the Store are split per sport or per data type - one for tennis, another for live scores, another for odds. This one covers every sport that carries data, in every mode, with a single input schema. Ask for sportIds: [1, 2, 3] and you get football, tennis and basketball in the same dataset with the same field names.
• A whole day in one request. fixtures mode returns every match Flashscore publishes for a sport on a given day - roughly 275 football matches and 470 tennis matches on an ordinary day - already grouped with their tournament and country.
• Real match analytics, not just the score. match mode returns the statistics panel as structured rows: expected goals (xG), ball possession, total shots, shots on target, big chances, corners, and for tennis aces and double faults. That is the data people actually model with.
• Past and future, not only today. dayOffsets walks roughly a week back and a week forward, so you can backfill last weekend's results and pull next weekend's fixture list in the same run.
• Nothing is silently dropped. Flashscore's feed uses undocumented two-letter keys. This Actor names the ones it has positively identified and ships the untouched upstream record alongside them under _raw, so a key that Flashscore adds tomorrow still reaches your dataset.
• Honest about gaps. Where a value could not be verified - the fine-grained status code, incident type IDs - it ships raw and is documented as raw instead of being guessed into a pretty label.

How it works

Flashscore's website is a JavaScript app: its HTML is ~740 KB of shell and bundles and contains no match rows. All data is loaded from a separate feed host, local-global.flashscore.ninja, which answers plain HTTP GET requests as long as they carry one static header that Flashscore's own bundle sends. There is no login, no cookie warm-up, no challenge and no rate-limit gate on the endpoints this Actor uses.

The feed does not return JSON. It returns Flashscore's own delimited encoding, where records are separated by ~, fields by ¬, and keys from values by ÷:

SA÷1¬~ZA÷ARGENTINA: Liga Profesional¬ZEE÷naYhNOaA¬ZY÷Argentina¬~AA÷WSSid6OM¬AD÷1788818400¬AE÷Barracas Central¬AF÷Argentinos Jrs¬AG÷0¬AH÷0¬...

The Actor decodes that into records, carries each tournament header forward onto the matches that follow it (the tournament is not repeated on the match rows), converts timestamps to ISO 8601, maps the status and winner codes, and adds a public match URL. Full details of how the format was mapped - including which keys appear on more than 90% of rows and which only appear once a match kicks off - are in CRAWLING_METHOD.md.

Input

Field	Type	Default	Description
mode	string	fixtures	fixtures, live, match or reference.
sportIds	array	["1"]	Sport IDs to scrape (see table below). In match mode this only labels the output.
dayOffsets	array	["0"]	fixtures mode. 0 today, -1 yesterday, 1 tomorrow; roughly -7..+7 is served.
matchId	string	-	match mode. One 8-character match ID (e.g. WSSid6OM) or a full Flashscore match URL.
matchIds	array	-	match mode. Several IDs or URLs in one run.
includeSummary	boolean	true	match mode. Goals, cards and substitutions with minute, player and side.
includeStatistics	boolean	true	match mode. Statistics panel including xG and possession for football.
includeLineups	boolean	false	match mode. Formation, starting XI, bench, player ratings when published.
includeHeadToHead	boolean	false	match mode. Previous meetings plus recent form for both sides.
maxItems	integer	200	Cap for fixtures/live, split evenly across every sport/day combination.
maxConcurrency	integer	4	How many match IDs to fetch in parallel in match mode.
proxyConfiguration	object	disabled	Optional. Not needed in normal use - see Notes & limits.

Sport IDs

ID	Sport	ID	Sport
1	Football	11	Futsal
2	Tennis	12	Volleyball
3	Basketball	13	Cricket
4	Ice Hockey	14	Darts
5	American Football	15	Snooker
6	Baseball	21	Badminton
7	Handball
10	Bandy

Run reference mode to see which IDs are carrying matches today, measured live rather than read from this table.

Example inputs

Every football match today

{ "mode": "fixtures", "sportIds": ["1"], "dayOffsets": ["0"], "maxItems": 300 }

Weekend backfill across three sports

{ "mode": "fixtures", "sportIds": ["1", "2", "3"], "dayOffsets": ["-2", "-1", "0"], "maxItems": 900 }

Only what is in play right now

{ "mode": "live", "sportIds": ["1", "3", "4"], "maxItems": 100 }

Deep dive on specific matches

{
  "mode": "match",
  "matchIds": ["WSSid6OM", "2myy0lPc"],
  "includeSummary": true,
  "includeStatistics": true,
  "includeLineups": true,
  "includeHeadToHead": true
}

Which sports are live today

{ "mode": "reference" }

Output

Every record carries the envelope fields _input, _source, _scrapedAt, plus _error/_errorDetail when something failed. recordType tells the three shapes apart: MATCH, MATCH_DETAIL and SPORT.

MATCH (fixtures and live modes)

{
  "_input": "fixtures:sport=1:day=0",
  "_source": "S1-feed",
  "_scrapedAt": "2026-09-08T12:28:04Z",
  "recordType": "MATCH",
  "sportId": 1,
  "sportName": "Football",
  "tournament": "ARGENTINA: Liga Profesional - Clausura",
  "country": "Argentina",
  "tournamentId": "naYhNOaA",
  "tournamentStageId": "lEqmiyVI",
  "matchId": "2myy0lPc",
  "matchUrl": "https://www.flashscore.com/match/2myy0lPc/",
  "homeTeam": "Union de Santa Fe",
  "awayTeam": "Instituto",
  "homeTeamId": "vRMEOMPn",
  "awayTeamId": "vJEG4Uhc",
  "homeScore": "3",
  "awayScore": "2",
  "score": "3-2",
  "status": "FINISHED",
  "statusCode": "3",
  "winner": "HOME",
  "startTime": "2026-09-08T00:15:00Z",
  "finishedAt": "2026-09-08T02:07:31Z",
  "homeTeamImage": "YVxbV1iT-KxCl9RDG.png",
  "_raw": { "AA": "2myy0lPc", "AD": "1788826500", "...": "every upstream key, untouched" }
}

MATCH_DETAIL (match mode)

{
  "recordType": "MATCH_DETAIL",
  "matchId": "WSSid6OM",
  "matchUrl": "https://www.flashscore.com/match/WSSid6OM/",
  "meta": { "DA": "3", "DC": "1788818400", "DD": "1788825065" },
  "statistics": [
    { "period": "Match", "statId": "432", "name": "Expected goals (xG)", "home": "0.37", "away": "0.29" },
    { "period": "Match", "statId": "12",  "name": "Ball possession",     "home": "42%",  "away": "58%"  },
    { "period": "Match", "statId": "34",  "name": "Total shots",         "home": "7",    "away": "7"    },
    { "period": "Match", "statId": "13",  "name": "Shots on target",     "home": "1",    "away": "3"    },
    { "period": "Match", "statId": "459", "name": "Big chances",         "home": "1",    "away": "1"    },
    { "period": "Match", "statId": "16",  "name": "Corner kicks",        "home": "3",    "away": "4"    }
  ],
  "incidents": [
    { "stage": "1st Half", "minute": "25'", "player": "Briasco N.", "side": "HOME", "incidentTypeId": "1" },
    { "stage": "1st Half", "minute": "33'", "player": "Prieto S.",  "side": "HOME", "incidentTypeId": "2" }
  ],
  "lineups": [ { "LA": "Formation", "LD": "1-5-3-2", "LRH": "6.9" } ],
  "headToHead": [ { "KB": "Last matches: Barracas Central" } ]
}

A finished football match typically yields around 120 statistics rows (per period and per stat), 10-20 incidents, and 50+ lineup blocks.

SPORT (reference mode)

{
  "recordType": "SPORT",
  "sportId": 2,
  "sportName": "Tennis",
  "isNamed": true,
  "matchesToday": 470,
  "sampleTournament": "ATP - SINGLES: US Open (USA), hard"
}

What you can build with it

• Live score feed for a bot, dashboard or notification service - run live mode on a schedule of a few minutes across the sports you care about.
• Results database - run fixtures daily with dayOffsets: ["-1"] and append; every row already carries tournament, country, both team IDs and the final score.
• Model inputs for football analytics - xG, possession, shots and big chances per match, joined on matchId from your fixtures table.
• Fixture calendar - dayOffsets: ["1","2","3"] gives you what is coming, with kick-off times already in UTC.
• Head-to-head previews - for each upcoming match ID, run match mode with includeHeadToHead and you have both sides' recent form.
• Cross-sport coverage monitoring - reference mode gives a daily count of matches per sport, useful to decide where the data is dense enough to build on.

Notes & limits

• Scores are absent, not zero, before kick-off. An unstarted match has no score keys at all. The Actor deliberately does not collapse that into 0-0; score simply is not present.
• status is coarse on purpose. SCHEDULED/LIVE/FINISHED are mapped from the status key that appears on more than 90% of rows. The finer code (values such as 4, 11, 12, 13, 38, 45 - covering half-time, extra time, postponed and similar states) is not documented upstream and was not exhaustively observed, so it ships untranslated as statusCode.
• Incident type IDs ship raw for the same reason. Goals and cards are distinguishable in practice, but the full ID map is not published, so this Actor does not pretend to know it.
• Lineups are not always published. Lower-division and youth matches often have none; the feed then returns an empty body and the field is omitted rather than faked.
• The live feed also carries the day's other matches. The Actor filters strictly to in-play rows, so a run during a quiet hour legitimately returns zero records.
• match mode ignores the sport you pass. The detail feed is keyed on the match ID alone - the same ID returns byte-identical data whatever sport number is sent - so sportIds there only labels the output record.
• Old matches disappear upstream. Flashscore drops matches from the feed host after a while; a long-gone ID returns _error: not_found rather than an empty record.
• No proxy needed in normal use. Every recon request succeeded on a direct connection. Proxy support is wired up for regions or IPs that get rate-limited.
• Standings and odds are not included. Neither could be located on the feed host with the endpoint names tried during recon; rather than ship a half-working mode, they are left out and documented here as a known gap.

Per-sport notes

Coverage is not uniform across sports, and it is worth knowing that before you build on it. Measured on an ordinary weekday:

Sport	Matches in a day	What the detail feed gives you
Football (1)	~275	The richest set: xG, possession, shots, big chances, corners, goals/cards by minute, lineups with ratings, H2H
Tennis (2)	~470	Aces, double faults, service and return stats per set; no lineups (individual sport)
Basketball (3)	~25	Score by quarter, team stats
Ice Hockey (4)	~40	Period scores, goals and penalties
Baseball (6)	~20	Innings, hits and errors
Handball (7), Volleyball (12)	10-15	Set/period scores
Cricket (13)	~10	Innings scores; formats vary by competition
Darts (14), Snooker (15), Badminton (21)	15-45	Frame/leg/set progress
American Football (5), Bandy (10), Futsal (11)	1-5	Thin - fine for scores, too sparse to model on

Tennis and football are where the volume is. If your use case needs one dense sport, start there; if it needs breadth, reference mode tells you every day which sports are actually carrying matches.

Scheduling tips

• Live tracking: run live mode every 2-5 minutes for the sports you care about. One run is one request per sport, so this stays cheap even at a high cadence.
• Daily results archive: run once after midnight UTC with dayOffsets: ["-1"] and append to the same dataset. matchId is a stable primary key for de-duplication.
• Fixture calendar: run weekly with dayOffsets: ["1","2","3","4","5","6","7"].
• Analytics pipeline: run fixtures first, filter to status: FINISHED, then feed those matchId values into a second match-mode run with includeStatistics. That two-step pattern keeps the expensive per-match calls to matches you actually want.

Troubleshooting / FAQ

I got zero records in live mode. That is a real answer, not a failure: nothing was in play in those sports at that moment. Football has quiet windows of several hours depending on the time zone.

A match has no score field. It has not kicked off. Scores are absent before the start rather than zero - see Notes & limits.

_error: not_found on a match ID that worked last month. Flashscore drops old matches from the feed host. Pull detail close to when the match is played, or archive it yourself.

_error: unauthorized. The static header the feed requires has been rotated upstream. Nothing you can configure; the Actor needs a one-line update. It raises immediately instead of retrying, so a run in this state costs almost nothing.

Lineups came back missing. Many lower-division, youth and women's matches never publish one. The Actor omits the field rather than inventing an empty structure.

Can I get league tables or betting odds? Not in this version - see Notes & limits. Both exist on the site, but the feed IDs for them were not found during recon and shipping a guess would be worse than shipping nothing.

Why is the tournament a single string like ENGLAND: Premier League? That is exactly how Flashscore publishes it. The country is split out separately into country for convenience, and tournamentId/tournamentStageId give you stable keys to group by.

Do I need a proxy? No. Every recon request succeeded on a direct connection, with no cookies and no warm-up. Proxy support is available if your own IP range gets rate-limited.

What does _raw contain? The untouched decoded record straight from Flashscore. If they add a key that this Actor does not name yet, you still receive it.

Legal & fair use

This Actor reads only data that Flashscore serves publicly to any visitor - no login, no paywall, no personal data. Sports fixtures, scores and statistics are facts, but Flashscore's compilation of them may be protected in some jurisdictions, and their terms of use apply to you as the operator. Use it for analysis, research and internal tooling; check your own legal position before republishing large extracts commercially. Keep run frequency reasonable - the whole point of the feed approach is that a day of matches costs one request, not thousands.