PYTH PRICE IN JPY FOR USD 24/7
The PYTH PRICE IN JPY FOR USD 24/7 (FX.Index.USD/JPY) price feed delivers a live USD price, updated in real time on Pyth Network. Prices are aggregated from first-party publishers - exchanges, market makers and trading firms pricing from their own order flow - with a confidence interval on every update. Available on 100+ blockchains.
FX is one of the few major asset classes with no central exchange, no consolidated tape and no official closing price. The BIS Triennial Survey put average daily turnover at roughly $9.5 trillion in April 2025, a 27% increase over 2022, of which spot was about $3.0 trillion; the rest is swaps, forwards and options. Because the market is over-the-counter and credit-intermediated, there is no single EUR/USD price. What exists is a distribution of executable quotes that differ by venue, counterparty tier, trade size and credit relationship: primary central limit order books for interbank majors, multi-dealer pools that carry much institutional buy-side flow, single-dealer bank platforms that internalise much of their client flow, non-bank market makers that supply a significant share of top-of-book liquidity in majors, and retail aggregators that keep quoting in thin periods.
Trading runs continuously from Sunday 5:00 PM ET to Friday 5:00 PM ET as liquidity passes between Sydney, Tokyo, London and New York. There is no daily maintenance break. Instead there is a liquidity hole at the 5:00 PM ET rollover, when value dates change, spreads widen and some venues briefly stop quoting, and a 48-hour weekend gap during which 24/7 perpetual-swap venues can serve as a reference for price discovery. London is the deepest session (around 38% of turnover books through UK desks, BIS April 2025) and the London/New York overlap carries the tightest spreads and most US data releases.
Pyth FX feeds aggregate first-party contributions from institutions that are themselves making prices in the currency market: the network receives the price each publisher is quoting, plus that publisher's own uncertainty. Every update carries a confidence interval, and applications pull the price on demand with a staleness threshold that matches the feed's schedule. Because FX has no exchange print to anchor to, the confidence interval is a direct expression of where the distribution of executable quotes sits: it typically widens at the rollover, at the Sunday open and outside local hours for emerging-market pairs, so the dispersion is part of the output rather than averaged away.
FX is one of the few major asset classes with no central exchange, no consolidated tape, and no official closing price. The BIS 2025 Triennial Survey put average daily turnover at roughly $9.5 trillion on a net-net basis in April 2025, a 27% increase over 2022 (BIS Quarterly Review, December 2025).
Because FX is over-the-counter (OTC) and bilaterally credit-intermediated, different counterparties have different prices even for the same FX pair. What exists instead is a distribution of executable quotes that differ by venue, by counterparty tier, by trade size, and by credit relationship. A tier-1 bank streaming to a hedge fund through its single-dealer platform and a retail aggregator streaming to a broker quote the same pair (e.g. EUR/USD) at similar levels, because arbitrage between venues keeps the distribution tight, but it never collapses to one number.
This differs fundamentally from oil, where a futures contract on NYMEX or ICE produces a single continuous, transparent, exchange-printed price series. Price discovery lives in the OTC market, not on an exchange.
| Tier | Participants | Role in price formation |
|---|---|---|
| Primary venues | Interdealer and exchange-group FX platforms | Central limit order books for interbank majors. Historically the reference for 'the' spot price. |
| ECNs / aggregators | Multi-dealer trading platforms and aggregators | Multi-dealer liquidity pools. A major execution channel for institutional buy-side flow. |
| Single-dealer platforms | Global banks streaming proprietary prices to clients | Internalised flow. Large banks internalise much of their client flow without hitting the primary market. |
| Non-bank market makers | Principal trading firms streaming to venues and clients |
| Factor | Impact |
|---|---|
| Rate differentials | One of the main medium-term drivers. FX is largely a relative-rates trade; changes in expected policy paths move spot more reliably than realised inflation or growth |
| Central bank policy | Scheduled decisions (Fed, ECB, BoJ, BoE) and unscheduled guidance. Unlike OPEC, most land during market hours. But BoJ and PBoC actions frequently land in thin Asian sessions |
| Data releases | US CPI and non-farm payrolls are among the highest-impact scheduled events |
| Intervention | Direct central bank action (MoF/BoJ in USD/JPY, SNB historically in EUR/CHF, PBoC via the daily fix) |
| Capital controls | For EM: onshore/offshore spreads (CNY vs CNH), NDF-versus-deliverable divergence, and administrative fixings that can override market-clearing levels |
| Terms of trade | Commodity exporters (AUD, CAD, NOK, IDR, CLP) track their export complex, creating cross-asset correlation between FX and commodity feeds |
Japan's Ministry of Finance bought yen against the US dollar on 30 July, 31 July and 3 August 2026, after the yen had fallen past 163 per dollar, its weakest level since 1986. The 31 July operation (US Eastern Time) was conducted in coordination with the US Department of the Treasury, per the Ministry's statement of 3 August 2026. The Ministry's published data put the July-September 2026 total at ¥15,399.3 billion, of which ¥7,975.6 billion was bought on 30 July (Ministry of Finance, 9 November 2026). For a feed consumer the lesson is the shape of the move: intervention lands without advance notice and can move the price sharply within minutes, so gate on confidence width as well as age.
In futures markets the key liquidity discontinuity is the contract roll. In FX there is no expiry on the spot price, so the liquidity structure is driven entirely by clock and calendar. Three windows matter.
Every trading day at 5:00 PM ET, the spot value date rolls forward and books are handed between New York and Asia-Pacific desks. Top-of-book depth in majors thins, spreads widen, and some venues briefly stop streaming. Prints during this window are real but unrepresentative: they reflect whoever was still quoting, not where size could actually trade.
The traditional FX market schedule closes Friday 5:00 PM ET and reopens Sunday 5:00 PM ET. Anything that happens in those 48 hours - elections, central bank emergency action, geopolitical escalation, sovereign downgrades - is expressed as a single discontinuous jump at the Sunday open, into the thinnest liquidity of the week.
EM pairs nominally trade around the clock, but genuine two-way liquidity is concentrated in local market hours. Outside those windows quotes become indicative, spreads widen materially, and confidence intervals should be expected to expand.
| Currency | Pyth market hours | Hours of most liquidity |
|---|---|---|
| USD/CNH (Offshore Yuan) | 6PM (prev. day) - 5PM ET | 8AM - 11PM CST |
| USD/INR (Indian Rupee) | 6PM (prev. day) - 5PM ET | 9AM - 5PM IST |
| USD/IDR (Indonesian Rupiah) | 6PM (prev. day) - 5PM ET | 8AM - 4PM WIB |
Pyth FX feeds aggregate first-party contributions from institutions that make prices in this market. Publishers are firms with direct market access, their own pricing engines and the ability to distribute their own data, including exchanges and trading venues, banks, non-bank market makers and trading firms. Pyth does not enumerate individual publishers per feed. Each publisher submits a price and a confidence interval on every update.
The Pyth price for each FX feed, including FX.EUR/USD, is an indicative reference price derived from an aggregate of publisher submissions. It is venue-agnostic: it does not represent the price available on any particular venue or from any dealer, and it is not an offer to buy or sell currency.
| Traditional FX data | Pyth Network |
|---|---|
| Redistributed venue feeds, per-terminal licensed | First-party contributions from the institutions that own the data |
| A single price with no uncertainty measure | Price plus confidence interval with every update |
| A single composite rate, so cross-venue dispersion is invisible | Aggregate across publisher types; when quotes disagree the confidence interval widens |
| Permissioned terminal access, not consumable by a smart contract | Same feed ID on 100+ blockchains plus REST and WebSocket APIs |
Pyth Pro (formerly Lazer) is the subscription-based, low-latency delivery layer for institutional and latency-sensitive consumers, delivered over standard WebSocket and REST APIs (FIX on request). Figures below are per Pyth's published documentation as of September 2026.
| Metric |
|---|
Always verify price freshness before use with
getPriceNoOlderThan()and a threshold appropriate to this feed's schedule. A threshold that is correct during the trading session will reject every price while the underlying market is closed; threshold logic must be schedule-aware, not a single global constant.
FX majors update continuously during the trading week, so a tight threshold is appropriate - but that same threshold will reject every price for 48 hours over the weekend and, for EM pairs, outside local hours (USD/BRL, for example, is scheduled 8:00 AM - 5:00 PM ET only). Read the schedule from feed metadata and derive the threshold per pair and per session.
| Use case | Recommended approach | Formula |
|---|---|---|
| Valuing collateral (lending) | Use lower bound | μ - σ |
| Outstanding loan positions | Use upper bound | μ + σ |
| Extreme volatility (σ/μ exceeds threshold) | Pause new activity for that feed | - |
| Rollover window and Sunday open | Widen thresholds or pause price-sensitive operations | - |
| EM FX outside local hours | Expect and tolerate wider bands; do not treat as an anomaly | - |
Quote convention and inversion. FX pairs are quoted by market convention, not alphabetically or by any rule a parser can infer. EUR/USD and GBP/USD are quoted with USD as the quote currency; USD/JPY, USD/CHF, and USD/CAD are quoted with USD as the base. Hardcoding an assumption that USD is always the quote currency is a common FX integration bug, and it produces a price that is wrong by a factor of the price squared rather than obviously broken.
FX symbology is simpler than futures because there is no expiry to encode, but the convention question makes it easy to get wrong.
FX.{BASE}/{QUOTE}
Where BASE and QUOTE are ISO 4217 codes ordered by market convention. The price is the number of units of QUOTE per one unit of BASE. The 24/7 index feeds use FX.Index.{BASE}/{QUOTE}.
| Symbol | Reads as | Convention note |
|---|---|---|
| FX.EUR/USD | USD per 1 EUR | EUR is base against everything |
| FX.GBP/USD | USD per 1 GBP | GBP is base against everything except EUR |
| FX.USD/JPY | JPY per 1 USD | USD is base - inverse orientation to EUR/USD |
| FX.USD/IDR | IDR per 1 USD | USD base for essentially all EM |
| FX.EUR/JPY | JPY per 1 EUR | Direct cross feed; do not triangulate if this exists |
Note: exponents differ by pair. USD/JPY and USD/IDR carry different scaling to EUR/USD because of the difference in quote magnitude - IDR quotes in the thousands, JPY in the hundreds, EUR/USD around 1. Never hardcode an exponent; read it from the feed.
All feed IDs verified against Hermes (https://hermes.pyth.network/v2/price_feeds?asset_type=fx) on 2026-11-11. Schedules are the Hermes schedule strings in America/New_York.
Sources:
https://hermes.pyth.network/v2/price_feeds?asset_type=fx - feed IDs and schedules (2026-11-11)Data vintage: 2025 (BIS turnover figures, April 2025) and 2026 (Pyth Pro figures per published documentation, September 2026; feed coverage and schedules, Hermes, 11 November 2026; Japan intervention data, Ministry of Finance, November 2026).
Links: feed page https://app.pyth.com/explore/FX.EUR%2FUSD; documentation https://docs.pyth.network.
This document is for informational and educational purposes only and does not constitute legal, financial, investment or other professional advice. Market-structure descriptions are provided for general context and are not a complete or authoritative statement of applicable law or market structure. Accuracy and performance figures reflect historical observations over specific periods and are not a guarantee of future performance or data quality. References to exchanges, venues, protocols or products are illustrative and do not constitute an endorsement. Digital assets and decentralised finance protocols carry significant risk, including total loss. Conduct independent due diligence and consult qualified professionals before acting on this information.
Pyth FX feeds are live spot exchange rates for currency pairs, symbolised FX.{BASE}/{QUOTE} (for example FX.EUR/USD). Each feed aggregates first-party price contributions from institutions that are themselves making prices in the currency market - venues, banks, non-bank market makers and trading firms. Every update carries a price and a confidence interval.
As of 11 November 2026 this pack covers seven majors, two major crosses, ten emerging-market pairs and three 24/7 index feeds (EUR/USD, USD/JPY and USD/BRL), available on 100+ blockchains per Pyth's documentation and through Pyth Pro's REST and WebSocket APIs. Because FX has no central exchange or official close, the feed represents a distribution of executable quotes, and the confidence interval is the measure of how wide that distribution is at any moment.
Pairs follow market convention, not a rule a parser can infer. EUR/USD and GBP/USD quote USD per unit of EUR or GBP; USD/JPY, USD/CHF, USD/CAD and essentially all emerging-market pairs quote the foreign currency per one USD. The price is always units of QUOTE per one unit of BASE. Assuming USD is always the quote currency is a common FX integration bug, and it produces a value wrong by a factor of the price squared rather than obviously broken.
Read base, quote and exponent from feed metadata. If you invert, invert the confidence interval too: for price p and interval s, the inverted interval is approximately s/p^2. For crosses, prefer a directly published feed such as FX.EUR/JPY over triangulation; if you must triangulate, propagate both intervals.
The traditional FX market closes Friday 5:00 PM ET and reopens Sunday 5:00 PM ET, and standard Pyth FX feeds follow that schedule. A feed that stops updating on Saturday is behaving correctly; the same stop at 10:00 AM on a Tuesday is an incident. Anything that happens during those 48 hours - elections, emergency central bank action, geopolitical events - is expressed as a single jump at the Sunday open, into the thinnest liquidity of the week.
| A significant share of top-of-book liquidity in majors. |
| Retail / CFD | Brokers, aggregators, prime-of-prime | Price takers. Wider spreads, but a visible source of continuous quotes in thin periods. |
The headline turnover figure is dominated by FX swaps, not spot. Average daily turnover by instrument, April 2025 (BIS Triennial Survey 2025, preliminary release, September 2025):
| Instrument | ADV (Apr 2025) |
|---|---|
| FX swaps | $4.0tn (42%) |
| Spot | $3.0tn (31%) |
| Outright forwards | $1.8tn (19%) |
| FX options | $0.7tn (7%) |
Pyth FX feeds are spot rates derived from institutions actively quoting and trading the pairs themselves.
FX runs continuously from Sunday 5:00 PM ET to Friday 5:00 PM ET as liquidity is handed off between Sydney, Tokyo, London, and New York. There is no daily maintenance break in majors, which distinguishes FX from the likes of commodity futures. Instead, there is a 2-day break plus a daily liquidity hole at the 5:00 PM ET rollover when books are passed between regions and value dates change.
With the rise of perpetual swaps, onchain and offchain venues that offer 24/7 trading on FX pairs can serve as a reference for price discovery during the 2-day weekend gap.
| Session | Approximate window (ET) | Character |
|---|---|---|
| Sydney / Wellington | 5:00 PM - 2:00 AM | Weekly open. Typically the thinnest liquidity of the week |
| Tokyo | 7:00 PM - 4:00 AM | JPY, AUD, NZD, Asian EM. Tokyo fix at 9:55 AM JST |
| London | 3:00 AM - 12:00 PM | Deepest liquidity globally. ~38% of turnover books through UK desks (BIS, April 2025) |
| London / NY overlap | 8:00 AM - 12:00 PM | Tightest spreads, largest volumes, most US data releases |
| New York | 8:00 AM - 5:00 PM | The 4:00 PM London fix falls in the overlap; NY afternoon liquidity thins |
| Daily rollover | 5:00 PM | Value date rolls; spreads widen and some venues briefly stop quoting |
| Weekend | Fri 5:00 PM - Sun 5:00 PM | Traditional venues closed; 48 hours of news expressed as a single jump at the Sunday open |
| USD/KRW (Korean Won) | 6PM (prev. day) - 5PM ET | 9AM - 3:30PM KST |
| USD/PHP (Philippine Peso) | 6PM (prev. day) - 5PM ET | 9AM - 4PM PHT |
| USD/TWD (Taiwan Dollar) | 6PM (prev. day) - 5PM ET | 9AM - 4PM TST |
| USD/MXN (Mexican Peso) | 6PM (prev. day) - 5PM ET | 1AM - 3:30PM CST |
| USD/TRY (Turkish Lira) | 6PM (prev. day) - 5PM ET | 10AM - 6PM TRT |
| USD/ZAR (South African Rand) | 6PM (prev. day) - 5PM ET | 9AM - 5PM SAST |
| USD/BRL (Brazilian Real) | 8AM - 5PM ET | 9AM - 6PM BRT |
Note: EM FX liquidity is also significantly limited at the start of the trading week and during local holidays, which shows up directly as wider confidence intervals rather than as missing data.
Through the primary drivers explained above, there have been many FX tail risk events over the years. Below are some cases and how they unfolded:
| Event, date | What happened |
|---|---|
| SNB de-peg, 15 Jan 2015 | The SNB discontinued its 1.20 EUR/CHF minimum exchange rate without advance notice. EUR/CHF traded well below parity intraday before recovering to around parity later that day. Liquidity vanished for minutes and several retail FX brokers became insolvent. |
| GBP flash crash, 7 Oct 2016 | Sterling fell about 9% against the dollar in early Asian trading before retracing most of the move (BIS Markets Committee, January 2017). The BIS review found no single cause, pointing to a confluence of factors including the time of day and thin liquidity. |
| JPY flash crash, 3 Jan 2019 | USD/JPY dropped from ~108.7 to ~104.9 in minutes during the Japanese New Year holiday, with AUD/JPY and TRY/JPY moving far more. Recovered most of the move within the hour. |
| USD/JPY intervention, Jul-Aug 2026 | Japan's Ministry of Finance bought yen on 30-31 July and 3 August, the 31 July operation in coordination with the US Treasury; see the What Moves the Price section. |
| Pyth Pro |
|---|
| Update frequency | Configurable channels: real-time, 50 ms, 200 ms, 1 s |
| End-to-end latency | Sub-100 ms |
| Uptime SLA | >99.99% |
| Coverage | 3,000+ feeds across FX, equities, commodities, crypto, indices and rates (September 2026) |
| Chain availability | 100+ blockchains plus off-chain APIs (REST, WebSocket) |
| Confidence intervals | Included with every update |
Pyth Indices and the FX weekend gap. Pyth Indices are 24/7 single-asset index products that aggregate onchain and offchain venues to provide a continuous reference for assets whose underlying markets close. For FX, FX.Index.EUR/USD, FX.Index.USD/JPY and FX.Index.USD/BRL are live on Hermes (feed IDs in the Symbols & Feed Metadata section, verified 2026-11-12). They are the continuous reference for products that must quote through the Friday-to-Sunday gap.
Weekend and session handling.
FX.Index.EUR/USD), not a staleness-tolerance setting.Users can select favourable historical prices within staleness constraints. Guard against this with strict
getPriceNoOlderThan()thresholds calibrated to the feed and the session, and gate on confidence width in addition to age.
In FX this is most exploitable in exactly the windows described in the Liquidity Profile & Tail Events section - the 5:00 PM ET rollover, the Sunday open and EM off-hours - where the spread between the best and worst available recent print is widest relative to normal conditions.
Access. Pyth Pro: WebSocket, REST — request a demo key at data@dourolabs.xyz.
| Symbol | Feed ID (hex) | Schedule | Note |
|---|---|---|---|
| FX.EUR/USD | 0xa995d00bb36a63cef7fd2c287dc105fc8f3d93779f062f09551b0af3e81ec30b | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Primary feed. Major. Min channel real_time |
| FX.GBP/USD | 0x84c2dde9633d93d1bcad84e7dc41c9d56578b7ec52fabedc1f335d673df0a7c1 | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major |
| FX.USD/JPY | 0xef2c98c804ba503c6a707e38be4dfbb16683775f195b091252bf24693042fd52 | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major. USD base |
| FX.USD/CHF | 0x0b1e3297e69f162877b577b0d6a47a0d63b2392bc8499e6540da4187a63e28f8 | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major. USD base |
| FX.AUD/USD | 0x67a6f93030420c1c9e3fe37c1ab6b77966af82f995944a9fefce357a22854a80 | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major |
| FX.USD/CAD | 0x3112b03a41c910ed446852aacf67118cb1bec67b2cd0b9a214c58cc0eaa2ecca | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major. USD base |
| FX.NZD/USD | 0x92eea8ba1b00078cdc2ef6f64f091f262e8c7d0576ee4677572f314ebfafa4c7 | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major |
| FX.EUR/GBP | 0xc349ff6087acab1c0c5442a9de0ea804239cc9fd09be8b1a93ffa0ed7f366d9c | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major cross |
| FX.EUR/JPY | 0xd8c874fa511b9838d094109f996890642421e462c3b29501a2560cecf82c2eb4 | Mon-Thu 24h; Fri 00:00-17:00; Sun 17:00-24:00; 24 and 31 Dec close 17:00 | Major cross. Direct feed; do not triangulate |
| FX.USD/CNH | 0xeef52e09c878ad41f6a81803e3640fe04dceea727de894edd4ea117e2e332e66 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms. Local hours in the Liquidity Profile & Tail Events section |
| FX.USD/INR | 0x0ac0f9a2886fc2dd708bc66cc2cea359052ce89d324f45d95fadbc6c4fcf1809 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/IDR | 0x6693afcd49878bbd622e46bd805e7177932cf6ab0b1c91b135d71151b9207433 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms. Quotes in the thousands |
| FX.USD/KRW | 0xe539120487c29b4defdf9a53d337316ea022a2688978a468f9efd847201be7e3 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/PHP | 0x2bda7f268b52bfbc3f2e124c31445247647350db313caadc6771e6299e0a68c9 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/TWD | 0x489f02f2f13584026d63fd397c80ed3b414a2820c4d43da0306fc007fcd5a8e0 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/MXN | 0xe13b1c1ffb32f34e1be9545583f01ef385fde7f42ee66049d30570dc866b77ca | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/TRY | 0x032a2eba1c2635bf973e95fb62b2c0705c1be2603b9572cc8d5edeaf8744e058 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/ZAR | 0x389d889017db82bf42141f23b61b8de938a4e2d156e36312175bebf797f493f1 | Mon-Thu 00:00-17:00 & 18:00-24:00; Fri 00:00-17:00; Sun 18:00-24:00; 24 and 31 Dec close 17:00 | EM. Min channel 200 ms |
| FX.USD/BRL | 0xd2db4dbf1aea74e0f666b0e8f73b9580d407f5e5cf931940b06dc633d7a95906 | Mon-Fri 08:00-17:00; closed 25 Dec and 1 Jan | EM, local hours only. Min channel 200 ms |
| FX.Index.EUR/USD | 0xc7feb6665b0906fd7db8872e556382c92fcca92dd45bfc7c7dd7dfaf159fa0e9 | 24/7 | Pyth Indices continuous benchmark. Min channel 50 ms |
| FX.Index.USD/JPY | 0x6667e9bef5c747731c889cdb84c58c3efdce680c69465f6ca6928fc7db9d98cc | 24/7 | Pyth Indices continuous benchmark. Min channel 50 ms |
| FX.Index.USD/BRL | 0x0fcfcd5f0639484be5779af095c4eb311a89e6bff7941b11aa62ed718d52d4ff | 24/7 | Pyth Indices continuous benchmark; covers the hours outside the FX.USD/BRL local session |
Do not carry Friday's close forward as a live price by loosening a staleness threshold. If a product must quote through the weekend, use an explicit continuous reference: FX.Index.EUR/USD, FX.Index.USD/JPY and FX.Index.USD/BRL are 24/7 Pyth Indices built from liquid onchain and offchain venues, including perpetual-swap markets that trade FX through the gap.
Emerging-market pairs such as USD/IDR or USD/KRW are scheduled nearly around the clock on Pyth, but genuine two-way liquidity is concentrated in local market hours (for example 8:00 AM - 4:00 PM WIB for the rupiah, 9:00 AM - 3:30 PM KST for the won). Outside those windows quotes become indicative and spreads widen materially, and the feed's confidence interval expands to match.
That widening is information, not a fault: treat it as expected, tolerate wider bands, and gate price-sensitive operations on confidence width rather than flagging an anomaly. The same effect appears at the start of the trading week and on local public holidays, so maintain a holiday calendar per currency rather than per feed. USD/BRL is scheduled for local hours only; FX.Index.USD/BRL is the 24/7 reference outside them.
Pyth FX feeds aggregate first-party contributions from institutions that make prices in the currency market, including exchanges and trading venues, banks, non-bank market makers and trading firms. Each publisher submits the price it is quoting together with its own uncertainty around it, and Pyth aggregates those into a single price and confidence interval.
Publishers are firms with direct market access, their own pricing engines and the ability to distribute their own data. Pyth does not enumerate individual publishers in feed documentation. The resulting price is an indicative reference price: it is not the price available on any particular venue and not an offer to trade.
Through Pyth Pro, updates are delivered on configurable channels - real-time, 50 ms, 200 ms or 1 s - over WebSocket and REST with sub-100 ms end-to-end latency, per Pyth's published documentation as of September 2026. Onchain, applications pull the latest price when they need it.
Integration follows three rules. First, verify freshness with getPriceNoOlderThan() using a threshold derived from the pair's schedule, not a single global constant, so the check is tight during the week and does not reject every price over the weekend. Second, gate on confidence width as well as age: widen thresholds or pause around the 5:00 PM ET rollover, the Sunday open and the 4:00 PM London fix. Third, read base, quote, exponent and schedule from feed metadata.
Request a Pyth Pro demo key at data@dourolabs.xyz.