What it is, in two halves
SEDA presents itself in two ways that are worth keeping apart. The commercial site sells four packaged products under the banner of powering markets that never close: composite data feeds, bundled feeds, pre-IPO data, and an outcome resolution layer. The documentation describes something more fundamental underneath them — a layer-one blockchain whose purpose is to run verifiable programs that fetch data, and to move that data into contracts or APIs.
Both descriptions are accurate, and the difference matters to a reader. If you are a venue or a product team, you are being sold a data and resolution service. If you are a developer, you are being handed a network where you write the oracle yourself — what data to pull, from which sources, filtered and aggregated how, delivered where.
Why it appears in this directory
Most entries here let you look at prediction markets. SEDA is one of the few that let you build one, and the only one in the directory whose job is to decide what happens when a market ends.
Its role splits in two directions. Inbound, it consumes prediction market data: the venues Polymarket and Kalshi are listed among the providers whose data is aggregated into SEDA feeds, which is how prediction market prices become inputs to products elsewhere. Outbound, it resolves outcomes: a builder defines a market and its resolution logic, and SEDA's infrastructure fetches the qualifying data and settles it programmatically.
What it does not do
One distinction prevents a common misunderstanding. SEDA does not resolve Polymarket's markets. Polymarket's own documentation is explicit that its resolution runs through the UMA optimistic oracle: anyone may propose an outcome by posting a bond — typically around $750 — after which a two-hour challenge period runs, and a dispute escalates to a token-holder vote that can take the timeline to somewhere between four and six days. Those resolution calls land on Polygon through adapter contracts that are visible on the public explorer, and they are Polymarket's own path to settlement.
SEDA offers an alternative resolution layer for markets you build yourself — the same category of function, a different dependency. Reading the two as one thing would be a mistake, and the directory record previously implied as much, which is why this page states it plainly.
The resolution layer: Outcomes
Outcomes is the product that matters most here. Its premise is that if you can define an outcome, the infrastructure can resolve it: a price above or below a level, whether a team won, whether a temperature threshold was breached, whether a governance vote passed. Four families of source are supported — public market data across equities, commodities, crypto and foreign exchange; private data such as corporate actions, earnings and bilateral quotes; real-world events including sports, weather, elections and policy decisions; and on-chain state such as protocol metrics or lending positions.
The vendor's own pages show the intent with demonstrations rather than prose: a resolver view where each example outcome carries the source that decided it and the milliseconds it took, and a status block listing a production version, a median resolution latency, a count of markets resolved over the past thirty days and an uptime figure. Those figures are worth reading with the caveats below, because SEDA's own disclaimer defines them narrowly — the latency claim, for instance, describes endpoint query behaviour rather than end-to-end settlement time.
Feeding data in: Signal and Bundles
Signal is the data product: continuously updated composite feeds for assets that trade around the clock, assembled from multiple providers and venues so that a market can keep pricing on a weekend, a holiday or outside trading hours. The mechanism is aggregation with failover and outlier detection, plus a configurable smoothing and session-awareness layer for instruments whose reference venues close.
Latency is sold in tiers rather than as a single promise, from milliseconds for the fastest tier through sub-second tiers intended for perpetual venues and liquidation engines, to a second and beyond for index products, dashboards and settlement. Bundles sit on top: instead of one asset, you define a weighted combination — a basket, an index, a spread — with automatic rebalancing. The published example pairs a broad equity benchmark with a perpetual venue's data, which is a reasonable illustration of what bundling is for.
How the infrastructure actually works
The mechanism is worth understanding because it explains both SEDA's flexibility and its cost model. Data is fetched by Oracle Programs: deterministic, verifiable units of code that pull from an endpoint, apply your filtering and aggregation logic, and return a structured result. They are compiled to WebAssembly, deployed to the network and identified by an ID. Any public endpoint can be queried directly; for private or paid sources, a Data Proxy holds the API credential on the provider's behalf, which is how a feed can be built on data that is not openly crawlable without exposing the key.
Delivery then splits. SEDA Fast serves results as an API and a WebSocket stream, which suits off-chain products that need continuous data. SEDA Core posts requests on chain and returns results to contracts, with prover contracts deployed per EVM network and a Hardhat integration for reading them. In practice, a prediction market builder might use the on-chain path for what must be trustless and the API path for what only needs to be fast.
The token
SEDA runs on its own token with a genuinely unusual mechanic: network usage burns supply. Executing Oracle Programs, delivering data and deploying programs all cost tokens, and a defined portion is permanently removed from circulation as part of protocol process rather than sent to a treasury. Staking secures the network — validators on the main chain, and separately the node operators that serve data proxies — and holders can delegate to validators through the standard wallet interface and vote on governance proposals. Burns, staking and deployment activity are tracked on the project's own explorer trends page and by third parties including DefiLlama.
Reading the numbers on its site
SEDA's marketing pages carry prominent statistics — volume supported, data feeds in production, uptime, tokens burned — and every one of them is footnoted to a content disclaimer that defines it narrowly. The same disclaimer states that the figures are provided for information only and that SEDA makes no representation about their accuracy. That is not a criticism; it is unusually candid, and it tells you exactly how to use them: as vendor-reported scale indicators, not as audited metrics. Two of them also disagree with themselves, which is documented in the method notes below.
Where it sits in this directory
Layer the directory and the picture is clean. Polygonscan shows you what already happened on chain. DefiLlama measures venues over time. OddsPapi gives you prices across venues. SEDA sits below all of them, in the position of a supplier: it is infrastructure that a prediction market platform would adopt in order to exist, and it is priced and documented as a business-to-business service rather than a tool you open in a browser.





































