Latency as a first-class metric
Every message is timestamped at ingress and egress so you can audit real-world staleness. No hand-wavy “milliseconds” marketing — just measurable numbers.
Ultra-low latency sports data
The feed your traders stop complaining about.
Olympus streams millisecond-level prices and in-play incidents for football and major global sports, engineered for bookmakers and professional syndicates who care about edge and execution.
- Direct, low-hop connectivity into core exchanges & providers
- Streaming WebSocket plus battle-tested REST for control planes
- Normalised fixtures, markets & selections across providers
Built by traders and quants. Designed for desks where 200 ms actually matters.
Why Olympus instead of yet-another feed?
Olympus is not a rebadged firehose. It is built as a low-latency, high-integrity view of the market, with the plumbing and observability serious trading operations expect.
Normalised markets & runners
Cross-bookmaker mapping for fixtures, markets and selections. Build your models once and deploy across providers without glue-code hell for each integration.
Battle-tested reliability
Redundant ingest pipelines, automatic replay and persistent streams keep your desk informed when a single upstream goes sideways mid in-play.
Streaming first. REST where it makes sense.
Connect once over WebSocket to receive prices, incidents and state changes in a compact, versioned schema. Use REST for onboarding, historical queries and reconciliation.
- Single authenticated channel per sport or multiplexed
- Back-pressure aware hints to avoid flooding your stack
- Deterministic sequencing for reconciling fills vs moves
WebSocket snapshot
JSON
{
"t": "price_update",
"sport": "soccer",
"event_id": "EPL-2025-08-17-ARS-MCI",
"market": "match_odds",
"selection": "home",
"price": 2.04,
"liquidity": 12450,
"ts": "2025-08-12T15:24:09.418Z"
}Real payloads are denser; this is a human-readable sketch of the schema.