> ## Documentation Index > Fetch the complete documentation index at: https://filament.getgalaxy.io/llms.txt > Use this file to discover all available pages before exploring further. # Live Tennis API source connector > Read tennis matches, players, tournaments, historical results, and rankings Connector ID `livetennisapi` The Live Tennis API source reads current match data, player and tournament reference data, and optional historical results and rankings. It uses an [HTTP manifest](/pages/connectors/building-a-connector/http-manifests) targeting API `v1` and is **alpha**. It has been checked against the official API reference and Filament's manifest validation. Live account validation is pending. ## Configuration | Field | Scope | Default | Description | | - | - | - | - | | `api_key` | Connection | | Required. Secret. Live Tennis API key. | | `match_status` | Connection | `live` | Status for `matches` and `match_details`. Choose `live`, `upcoming`, `completed`, or `cancelled`. | | `ranking_system` | Connection | `atp` | Official ranking table. Choose `atp`, `wta`, `itf_jt`, `itf_mt`, `itf_wt`, `atp_doubles`, or `wta_doubles`. | Get a key from your [Live Tennis API account](https://livetennisapi.com/account). Connection testing reads `/usage`, which is authenticated and quota-exempt. See the [official API reference](https://docs.livetennisapi.com/reference.html) for authentication and plan entitlements. Live and upcoming matches, fixtures, players, tournaments, and usage are available with a free key. Completed and cancelled matches and the history resources need BASIC or an eligible Historical Data API plan. The limited free history allowance is unsuitable for backfills. Rankings need PRO or higher. Match detail includes market and analysis fields only when your plan allows them. ## Resources | Resource | Endpoint | Parent | | - | - | - | | `usage` | `GET /api/public/v1/usage` | | | `matches` | `GET /api/public/v1/matches` | | | `match_details` | `GET /api/public/v1/matches/{matchId}` | `matches` | | `fixtures` | `GET /api/public/v1/fixtures` | | | `players` | `GET /api/public/v1/players` | | | `tournaments` | `GET /api/public/v1/tournaments` | | | `history_matches` | `GET /api/public/v1/history/matches` | | | `history_tapes` | `GET /api/public/v1/history/matches/{matchId}` | `history_matches` | | `history_coverage` | `GET /api/public/v1/history/coverage` | | | `history_incidents` | `GET /api/public/v1/history/incidents` | | | `archive_matches` | `GET /api/public/v1/history/archive/matches` | | | `archive_match_details` | `GET /api/public/v1/history/archive/matches/{archiveId}` | `archive_matches` | | `archive_players` | `GET /api/public/v1/history/archive/players` | | | `rankings` | `GET /api/public/v1/rankings` | | Only `matches` and `fixtures` are selected by default. Reference tables, detail reads, and paid resources are optional. Selecting a child also reads its parent, even if the parent is not selected for writing. Matches and fixtures each emit one row per returned record. Matches follow `match_status`, while fixtures list upcoming scheduled fixtures. Players and tournaments emit one row per catalog entry. Tournament IDs remain stable across seasons. Usage emits one row for the calling key, including its daily history. History matches cover completed matches from January 2023 onward. Each history tape is one row per match with the score sequence, profiles, and coverage metadata kept as JSON. The default raw sequence is requested. It may be incomplete or include reconstruction with null timestamps and model values. Read the tape's `meta` before using it for analysis. Archive matches cover results from 1968 through 2022 in a separate ID space. `event_date` is the tournament start date. Archive detail adds serve statistics where recorded. Archive players use the composite key `tour`, `id`, which joins the archive result's player IDs within the same tour. These are not roster IDs. Rankings emit the latest published table for the selected system. They have no primary key because players outside the roster have null IDs and ranks can tie. History coverage is also keyless and emits one rollup object. History incidents emit one row per published incident, including correction and affected-match metadata. Incident exports are not downloaded. Nested data stays in JSON columns. Each resource has a trailing `raw` column for fields not projected separately. Match detail's plan-dependent market and analysis objects remain in `raw`. Dedicated event, stoppage, status-transition, withdrawal, live point, price-tick, statistics, rally, and charting feeds are outside this version's scope. So are archive reconstructed tapes, career and head-to-head queries, Elo and per-player as-of rankings, bulk packages, WebSocket feeds, broadcast overlays, and webhooks. Player and tournament detail lookups duplicate the catalog schemas and are not separate resources. ## Modes All resources support full reads. Use full replace for keyless rankings and history coverage. Full replace also removes matches that leave the selected status. Full upsert refreshes keyed records but retains rows no longer returned. Incremental reads are not supported. The upstream match change feed requires a server watermark and a cursor that replaces the initial filter. This connector uses ordinary full listings. Historical tapes and results may receive later corrections, so play dates are not used as update cursors. ## Behavior * **Auth** uses the `X-API-Key` header on every request. * **Pagination** reads list records from `data` in pages of 50, advances `offset` by 50, and stops when `meta.has_more` is false. It can continue through short or empty pages. Detail objects, usage, coverage, and incidents are unpaginated. * **Rate limiting** allows one request every two seconds with a burst of one, matching the free plan's 30 requests per minute. The HTTP engine retries throttled requests and honors `Retry-After`. Daily allowances are shared with other clients and are not tracked by Filament. * **Backfills** request unbounded listings without date filters. Player catalogs, archives, and per-match tapes can exceed daily quotas. Select these resources deliberately and use a plan with sufficient allowance. * **Consistency** reflects the API while pages are fetched. Offset reads are not atomic snapshots, and live matches can enter or leave the listing during a run. * **Errors** remain visible. Missing entitlement and unavailable detail responses fail the read rather than silently dropping records. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.