Goldsky

Goldsky for Trading

An onchain data provider for prediction markets, kept current for you

An onchain data provider for prediction markets curates fills, positions, and PnL from venues like Polymarket into a structured dataset, instead of leaving each team to write and maintain its own decoder. Goldsky streams order fills, matched trades, balances, and per-user PnL into Postgres, ClickHouse, or a webhook, with full history included from day one. When a venue changes contracts, Goldsky updates the dataset and the pipeline keeps running, so a desk finds out through a changelog, not a dashboard gone flat.

Talk to our team

Highlights

Order fills, taker matches, balances and PnL, ready to query
Live activity and complete history from the same source
Delivered into the database or warehouse your team already uses
Dataset updates handled for you when venues change contracts

Trusted by leading trading and blockchain teams

Curated prediction market data, maintained for you

Polymarket alone processes millions of trades on Polygon, and it is one of several venues worth watching. We keep the datasets current through venue changes, so your team queries trading activity rather than maintaining the pipeline that produces it.

Datasets kept current for you

Venue contract upgrades and new markets are absorbed into the curated datasets, so your team picks them up without rewriting decoders.

Answers your desk can act on

Open interest, volume, order flow, market-maker activity and per-user PnL are all queryable in the database your team already uses.

One source everyone trusts

Trading, analytics and risk read the same replicated history instead of reconciling three different scrapes of the same source.

The data you get

Curated datasets covering trading activity and positions, streamed to wherever you need them.

Order fills, at full granularity

Every partial or complete fill, with price, size in USDC and shares, side, fee, counterparty, and whether the order was maker or taker.

Matched orders

Taker-level matches give you a higher-level view of trading activity when you want one row per trade rather than per fill.

User balances and positions

Current outcome token positions per user, plus a positions dataset carrying PnL detail including average price and realized PnL.

Real time plus full history

Streams arrive as trades happen and include all historical data, so backtests and live dashboards read from the same source.

Alerts on what you care about

Filter the stream down to specific markets, wallets or trading patterns and push matches to a webhook the moment they occur.

Act on it automatically

Route filtered fills into a Compose task to mirror trades, rebalance or redeem winning positions, with durable retries and gas sponsorship.

Building it yourself vs. a maintained dataset

Table 1. Building a prediction-market indexer in-house vs. a maintained dataset.
Concern	Building it yourself	With Goldsky
Venue contract upgrades	A rewrite of your decoders and a gap in history each time	Datasets updated for you. Your pipeline keeps running
Data landed	Whatever you build a sink for	Postgres, ClickHouse, Kafka, Pub/Sub, object storage, or a webhook, several from one pipeline
History	Only from when you started indexing	Full history included from the start
Ownership	Yours, but yours to maintain	Every row delivered into your own database, not held in an account you query through a vendor

Built for regulated environments

Controls, documentation and support your reviewers expect, ready before they ask.

SOC 2 Type II

Independently audited controls, with reports available for your security review.

Dedicated infrastructure

Isolated compute and single-tenant options for your production workloads.

Granular access controls

Role-based permissions and audit logging across your whole organization.

Full observability

Monitoring, alerting and data-lineage visibility your auditors can follow.

Data residency

Cloud, on-premise or co-located, so data sits where your obligations require.

Regulation-ready data

Structured for MiCA, FATF Travel Rule and BSA/AML/KYC, with retention you set.

Deliver on your timeline

Banks, payment platforms and trading firms use Goldsky to get onchain products live in weeks. Our engineers work alongside yours from the first call through production.

Scoping call

We map your use case, the chains involved and the systems the data lands in, then give you a delivery estimate.

Technical deep-dive

Our engineers work with yours to confirm the design and the integration points.

Kickoff and go live

We stand up your integration and validate the data against your systems, ready for your team to build on.

Ongoing support

24/7 support in a shared channel with the engineers who run your infrastructure.

Scoping call

We map your use case, the chains involved and the systems the data lands in, then give you a delivery estimate.

Technical deep-dive

Our engineers work with yours to confirm the design and the integration points.

Kickoff and go live

We stand up your integration and validate the data against your systems, ready for your team to build on.

Ongoing support

24/7 support in a shared channel with the engineers who run your infrastructure.

Book a scoping call

Ready to skip the indexer?

Bring your use case to a 30-minute call and leave with a concrete plan and timeline.

Talk to our team

Frequently asked questions

What is an onchain data provider for prediction markets?

A service that curates order fills, positions, and PnL from prediction-market venues into a structured, queryable dataset, and maintains it through contract upgrades, instead of a team building its own indexer.

Can we still use the public Polymarket subgraphs?

No. Polymarket moved to v2 contracts in April 2026 and is not using subgraphs going forward, so public subgraph endpoints return incomplete or incorrect data. Turbo pipelines with v2 datasets are current.

What happens the next time the venue changes its contracts?

Goldsky updates the datasets and the pipeline keeps running. The v2 migration cost teams a decoder rewrite and a gap in history. Your desk finds out through a changelog instead.

Where can the data land?

Postgres and ClickHouse for direct querying, Kafka, Pub/Sub, or object storage for distribution, or a webhook to react to individual events. Several sinks can run from one pipeline.

Can we trade on it, not just analyze it?

Yes. A common pattern filters fills to a watched set of wallets and webhooks each one into a Compose task that signs and submits its own order, with durable retries and gas sponsorship.

Tokenized equities and RWA Turbo streaming pipelines Compose Polymarket chain page