What SEDA is today
SEDA has moved a long way from the early framing of publishing Polymarket's data on-chain. It now presents itself as oracle infrastructure for 24/7 markets: one integration that lets a protocol list equities, commodities, indices, pre-IPO valuations or event outcomes — around the clock, including weekends and holidays when the underlying venues are shut — and have them resolved automatically. Protocols including Hyperliquid, Injective, Dreamcash, Perps.Fun, Nunchi and Outcome are listed as users of that infrastructure.
That makes SEDA a different sort of directory entry: it is not a place anyone trades, and it never holds user funds. It is the layer under other people's markets — the thing that decides what the price was and who won.
The four products it sells
Signal — white-label composite price feeds for any asset, aggregating sources such as Pyth, Hyperliquid, Coinbase, Kraken, dxFeed, Finage, Blocksize, CoinGecko, CoinMarketCap and, notably for this directory, Polymarket, Kalshi and PM Insights. Feed quality options include automatic failover and outlier detection, with four latency tiers: Ultra at 5ms (with optional co-location, aimed at HFT), Fast at 200ms (perpetual exchanges, liquidation engines, mark prices), Standard at 500ms (DeFi protocols, lending, options pricing) and Core at 1s+ (indices, dashboards, analytics, settlement).
Bundles — composite feeds that combine several underlying assets into one instrument, the example given being a USA500 feed used on Hyperliquid, with the company citing $7B+ in volume supported.
Outcomes — the resolution layer that matters most for prediction markets: define an outcome in any verifiable source and SEDA resolves it, with a stated median resolution latency of 150ms (the site elsewhere quotes under 200ms) and, on its own dashboard, 1,408 markets resolved in the last 30 days. Resolution sources span public markets (equities, commodities, crypto, forex), private data a public crawl cannot see (corporate actions, earnings, private valuations, bilateral OTC quotes), real-world events (sports, weather, elections, policy) and on-chain state (DeFi metrics, governance votes, lending LTV). Advertised use cases run from 24/7 up/down markets and prediction markets to range/barrier exotics, autocallables and weather or commodity parametric triggers.
Pre-IPO — tick-by-tick data on private companies, with 5,000+ covered, sourced through partner Caplight — the piece that makes private-equity and IPO-event markets possible.
How it is built
SEDA is its own proof-of-stake Layer 1, described in its docs as a read-write network for deploying and accessing Oracle Programs — deterministic, verifiable compute modules that fetch external data, apply custom logic and return structured output — and Data Proxies, middleware that holds an API key with access to a given provider. Because on-chain execution is limited by block times, latency-sensitive access runs through SEDA FAST, an optimised off-chain execution layer. Developers get one endpoint to call from any chain rather than one integration per venue, which is the whole commercial idea.
The token, honestly
SEDA's token economics are tied to usage rather than to a venue: data requests, computation, data delivery and the deployment of Oracle Programs and Data Proxies all incur fees, and protocol rules burn part of the supply as that activity happens (the team has published burns of about 10m SEDA on Base). Validators and overlay node operators must stake, holders can delegate through Keplr, and governance runs through on-chain proposals. Supply is inflationary with no cap: about 759.5m of 1.023B total supply circulates, and the token trades around $0.0216 for a market cap near $16.4m — roughly 92% below its all-time high of $0.2636, though up over the past year.
What to weigh before relying on it
SEDA sells to businesses, not consumers: the path is a trial then a contract, and terms allow cancellation only with 30 days' written notice. Its terms are unusually explicit that it is neutral infrastructure — it does not facilitate, process or settle financial transactions, is not a financial intermediary, exchange or custodian, and never takes custody of assets. The corollary is the caveat that matters most here: the headline numbers are self-reported. Its own content disclaimer says the $19.5B+ "trading volume supported" is informational only with no accuracy representation, that "3.5M production data feeds" may include inactive or test instances, and that latency figures describe endpoint query behaviour and are always subject to a separate service-level agreement. For a venue deciding whether to outsource resolution, that is the difference between a marketing number and a contract.





































