ESPN Scores Scraper - NFL NBA MLB NHL Soccer Live Scores

What does ESPN Scores Scraper do?

Get a bounded snapshot of upcoming, live and final games from ESPN scoreboards. Each unique game becomes one flat dataset row with teams, scores, status, venue and UTC timestamps. This is a scoreboard snapshot, not a continuous live feed or a historical completeness guarantee.

How to scrape ESPN scores

1. Start with the default MLB scoreboard and five games.
2. Optionally select leagues such as football/nfl, basketball/nba, hockey/nhl or soccer/eng.1.
3. Set an optional calendar date, run the Actor, then inspect both the dataset and RUN_SUMMARY coverage.

API input {} uses the same small default. For a reproducible historical example:

{"mode":"scores","leagues":["baseball/mlb"],"date":"20260904","maxItems":5}

Use the API tab for integration examples. You can schedule independent snapshots and download the dataset as JSON, CSV, HTML or Excel. Game IDs help downstream systems update scores between snapshots.

Input and coverage

mode supports scores only. leagues accepts 1–20 league codes; duplicate codes are fetched once. The existing API shorthand of one league string remains supported. maxItems is a strict integer from 1 to 2000, default 5, and applies globally across all leagues, including rows already present after resume.

date must be a real calendar date in YYYYMMDD format. ESPN determines scoreboard calendar semantics; omission uses ESPN's default day, not your machine's local timezone. Output game timestamps are normalized to UTC. League schedules and upstream availability vary; one scoreboard response per selected league is requested, without a full-season pagination guarantee. Reaching the global limit can leave later leagues unrequested.

Output

Existing fields remain: league, event_id, name, short_name, date, state, completed, status_detail, home_team, home_abbr, home_score, home_record, away_team, away_abbr, away_score, away_record, venue, city, source, scraped_at. source_url links to the source scoreboard. Optional values can be null when ESPN omits them; pregame scores should not be interpreted as final scores.

Illustrative abbreviated row:

{"league":"baseball/mlb","event_id":"example-id","home_team":"Home team","away_team":"Away team","home_score":3,"away_score":2,"state":"post","completed":true,"source":"espn"}

Errors and empty days

An explicit empty events array is a valid no-games day: zero result rows and a no_games coverage entry. Transport errors, HTTP errors, malformed payloads and incomplete required game fields fail the run. Previously delivered rows remain available, and RUN_SUMMARY marks source_error; do not assume failed runs have complete league coverage. Each request has a 15-second timeout and no automatic retry loop.

The default dataset is authoritative on resume. Existing event IDs are skipped across leagues and resumes within the same run; separate new runs intentionally produce fresh snapshots. Output writes use the SDK's automatic dataset pricing event, without a duplicate manual charge. Budgets are checked before each source fetch and each write.

Pricing

Current configured price: $0.002 per delivered game ($2 per 1,000), plus the Actor Start event at $0.00005 × max(1, memory in GB). At 512 MB, illustrative event totals are $0.01005 for five rows, $2.00005 for 1,000 rows, and $0.30150 for 30 separate five-row runs. These are event-price calculations, not measured infrastructure costs; actual settings, tier and prices are shown before starting. Empty runs have no result charges but retain the start event. A result budget does not replace a runtime or platform usage limit.

Support

Report reproducible issues through the Actor Issues tab with league, date and run ID. This Actor is independent of ESPN and relies on ESPN's public endpoint remaining available. Review source terms for your use and keep polling frequency appropriate.