Pyth Network $PYTH

$PYTH Answering @wolfur's question below with a test. I build small Python market dashboards, so today I tried both paths for a free once-a-day price.

1. Hermes: /v2/price_feeds still answers without a key, but /v2/updates/price/latest returns 401 with a plain-text "unauthorized", as described.

2. On-chain push feed: I read the sponsored SOL/USD account (7UVimf…jLiE) straight from the public Solana RPC with a short standard-library Python script: one getAccountInfo call, then decode the PriceUpdateV2 layout. No key, no SDK. Result: SOL/USD 119.80 USD ± 0.01.

What I noticed: the price was already 26 s old when I read it, so the sponsored feed doesn't refresh every slot. Fine for a daily script, not for a perp.

Suggestion: the 401 body could say "API key required" and link the sponsored push-feed accounts per chain with a minimal read example. That one link would have saved me the detour.

Question for the team: which Solana feeds are guaranteed to stay sponsored, and what update cadence or deviation threshold do they follow?

$PYTH

I trade XAUUSD, so I went through Pyth Terminal's two gold feeds: XAU/USD (spot, "Regular" coverage) and the GOLD 24/7 Pyth Index.

1. Holidays are hidden in the UI. The XAU/USD "Data Coverage" panel only shows the weekly schedule (Mon-Thu 00:00-17:00 & 18:00-24:00 NY, Fri to 17:00, Sun from 18:00). But the feed's JSON has the real exceptions: 26 Nov closes 14:30, 27 Nov closes 14:45 with no evening session, 24 Dec closes 13:45, 25 Dec and 1 Jan closed, 26 Mar closed. A perp or lending protocol reading only the panel would miss these. Suggestion: add an "Upcoming session changes" row under the schedule.

2. The GOLD 24/7 index is marked always open with min_publishers = 1, and the page shows a disclaimer but no methodology link. Questions for the team:

- When spot gold is closed (Fri 17:00 to Sun 18:00 NY), what inputs set the weekend price?

- At the Sunday reopen, does the index jump to XAU/USD, or converge over time? Weekend gaps are exactly where leveraged gold positions get liquidated.

- What does min_publishers = 1 mean here: one data source, or one aggregator?

A short methodology page linked from the feed would make this index much easier to trust for gold perps on Solana.

$PYTH One interesting aspect of Pyth Network is that its price feeds include a confidence interval, not just a price. This matters for DeFi lending and liquidation systems, where uncertainty in price data can affect risk decisions.

A question for the team: what safeguards do you recommend when a confidence interval becomes unusually wide relative to the reported price? A practical example of how developers should handle this scenario would be valuable for protocols integrating Pyth.

$PYTH is a gem!

What stands out to me about $PYTH Network is how it makes market price data accessible to developers building on Solana and other blockchains. I explored the SOL/USD feed in $PYTH H Terminal and found the feed metadata, such as the feed ID, exponent, and minimum publisher count, useful for understanding how price feeds are identified and consumed.

The documentation also provides guides and code examples that can help developers get started. One thing I’d love to see is more beginner-friendly, end-to-end examples showing how to take a price feed and use it in a practical application.

What real-world application would you most like to see developers build with $PYTH price feeds?

$PYTH — tried the Hermes price API as a developer, with curl and no API key (Oct 2026). Three things stood out:

1. Mixed auth with a silent failure. Feed discovery still works anonymously: GET hermes.pyth.network/v2/price_feeds?query=SOL returns results. But the next call, /v2/updates/price/latest for SOL/USD, returns HTTP 401 with a plain-text body that just says "unauthorized". No JSON, no mention of the API key, no link. Since discovery works, a dev following an older tutorial assumes the endpoint is fine and their request is wrong. A JSON error like {"error":"API key required since Aug 26, 2026","docs":"..."} would save a lot of debugging.

2. Search is substring-based. query=SOL returns 25 feeds, including Crypto.RESOLV/USD, Equity.US.SOLV/USD, Equity.US.SOLS/USD and funding-rate feeds. Adding asset_type=crypto narrows it to 9, but RESOLV is still there. An exact-symbol option (or ranking the exact base symbol first) would make picking the right feed ID safer, which matters a lot for an oracle.

3. Old docs links lose context. docs.pyth.network/price-feeds/core/api-instances-and-providers/hermes, which many tutorials link to, now lands on the generic API Reference page instead of a Hermes page explaining the new pyth.dourolabs.app/hermes base URL and key setup.

Question for the team: will the legacy hermes.pyth.network host keep serving the unauthenticated /price_feeds route, or is that going behind the key too?

Scope: public API via curl + docs, Oct 11, 2026. No API key used.

$PYTH is a gem!

$PYTH is a gem!

$PYTH: A Solana Starter Example Would Make the Docs Easier to Use

I explored Pyth's live feeds on my phone and reviewed the Price Feeds documentation as a Rust/TypeScript developer.

What worked: The documentation clearly separates Pyth Core and Pyth Pro, with separate Getting Started sections.

What slowed me down: SOL/USD displayed price information and feed metadata, including Feed ID 6. When I searched USD/NGN, no results came up. Finding the right feed ID for a specific pair could be more intuitive for new developers.

Suggestion: Add an end-to-end Solana example showing how to find a feed ID, fetch a price, interpret its confidence value, and integrate it into a small application.

Question: Is a beginner-friendly Solana tutorial planned, and how should applications handle unsupported currency pairs like USD/NGN?

$PYTH SDK onboarding gap: the official hermes-client README still teaches the keyless setup that now returns 401.

What I tried (4 Oct 2026): copied the Quickstart from pyth-crosschain apps/hermes/client/js/README.md — new HermesClient("https://hermes.pyth.network", {}) — and also hit the REST endpoint directly:

curl -i "https://hermes.pyth.network/v2/updates/price/latest?ids[]=0xef0d8b6f...b56d"

-> HTTP 401, body "unauthorized".

The docs banner explains it (Hermes requires an API key since the Aug 26, 2026 Pyth Core upgrade), but the README's Quickstart (line 26), the example-runner command, and the "Pyth offers a free public endpoint at hermes.pyth.network" line (line 93) never mention auth. The SDK already supports keys: HermesClientConfig has accessToken, sent as Authorization: Bearer (and as ACCESS_TOKEN for SSE/WebSocket) in hermes-client@3.1.0.

Suggestion: update the Quickstart to new HermesClient(endpoint, { accessToken: process.env.PYTH_API_KEY }), add the key-signup link beside it, and fix the "free public endpoint" sentence + example command. Right now a builder copying the README hits a 401 that looks like an outage.

Question: should SDK users switch the base URL to the upgraded pyth.dourolabs.app/hermes, or does hermes.pyth.network stay valid once a key is set?

$PYTH volume steady but less noisy than meme coins.

Infra tokens move differently.

Interesting to compare sentiment vs actual transaction flow here.

$PYTH is a gem!

$PYTH: the Price Feeds page shows two different sets of headline numbers

I reviewed pyth.network/price-feeds today (13 Oct 2026). The page has two "Global Price Layer" stat blocks, both labeled "Figures as of September 2026":

- One says 3,600+ Price Feeds, 1,000+ Partners, $5T+ Volume Secured

- The other says 3,059+ Price Feeds, 1,000+ Partners, $2.3T+ Volume Secured

The Asset Classes tiles on the same page add up to 3,503 feeds (189 commodities, 827 crypto, 55 indices, 15 funding rates, 2,022 equities, 299 FX, 44 metals, 12 NAV, 40 rates). That is close to 3,600 but well above 3,059.

This looks like a layout variant (desktop vs mobile, for example) that was never updated. But "volume secured" is the number institutions and builders quote, so showing both $5T+ and $2.3T+ on one page hurts trust in an oracle whose product is accurate data.

Suggestion: pull all headline stats from one source, ideally the live figures on kpi.pyth.network, and link each one to that dashboard so readers can check them.

Question for the team: which figure is current, and does "volume secured" count cumulative or trailing volume?

Scope: public website review only. I did no integration, staking or trading, and I'm not affiliated with Pyth. Prepared with AI assistance.

I tested the $PYTH Pyth Playground with BTC/USD and SOL/USD using the shared demo-token flow.

The API Key field says it can be left empty to use a rate-limited demo token, but pressing Run with the field empty returns:

“Invalid API key. Check the key you pasted and try again.”

I reproduced this twice, including after a hard refresh.

For a new developer this is confusing because the UI explicitly suggests that no API key is required for testing. It would be better either to fix the demo-token fallback or show a clearer message if the shared demo token is temporarily unavailable or rate-limited.

$PYTH

Pyth Network — data is great, onboarding is not

Tested: Pyth Terminal (SOL/USD, 1D/1W), Code panel, Feed ID copy, Solana pull-integration docs, Hermes API key flow.

What works: Terminal loads fast, charts are clean, Feed ID is easy to grab, docs have copy-paste Solana examples that actually compile.

What doesn't:

1. Hermes API key is a wall — I hit "create API key" with zero explanation of why I need it, what the free tier gives, or what happens without it. Felt like a paywall before I understood the product.

2. SendUSD live demo from the docs — couldn't find it. Docs reference it, link is dead or buried.

3. No "Copy as fetch (JS)" in the Code panel. I wanted to test a price feed in 30 seconds from console — had to hand-write the fetch from docs instead.

Suggestions:

- Add rate limit info next to the API key creation (requests/sec, what counts).

- Put an inline pull vs push decision tree at the top of the Solana integration page — I wasn't sure which one I needed for 15 minutes.

- Fix or remove the SendUSD demo link.

Verdict: solid oracle, rough edges in first-run experience. The data is trustworthy; the onboarding makes you work for it.

I price prediction markets that settle on $PYTH spot, so today I went to pull XAU/USD and SOL/USD from Hermes at a few weekend timestamps to see how the metals feed behaves while COMEX is closed. I never got that far. hermes.pyth.network, hermes-beta, benchmarks.pyth.network and the upgraded pyth.dourolabs.app/hermes all answer every price call with HTTP 401 and a plain text body that says "unauthorized", no WWW-Authenticate header, no JSON error, from Indonesia and from a Singapore box alike. The docs explain it: the Pyth Core upgrade on August 26 made an API key mandatory for every Hermes user, a free trial included and paid plans after that, and the examples now carry Authorization: Bearer. Two things for the team. /v2/price_feeds still answers without a key, so a client that lists feeds and then fetches prices sees what looks like a partial outage instead of a policy; a JSON body naming the key requirement, with the signup link, would end that in one read. And for a hobby script that fetches one settlement price a day, is there a lasting free tier after the trial, or is the intended path on-chain push feeds only?

$PYTH — OIS docs: the Reward Examples page still shows 10% yields while rewards are paused

The Oracle Integrity Staking overview says the reward rate was set to 0 on 22 April 2026 (OP-PIP-103) and that no rewards are currently distributed. The Mathematical Representation page repeats it: y = 0, "stake is still at risk of slashing".

The Reward Examples page, one click away, carries no such notice. All six examples assume "Annual Rate of Rewards r = 10%" (Example 4: publisher 10.3% APY, delegators 9.8%), and the built-in calculator returns positive rewards. Example 5 shows the other side: a 5% maximum slashing rate, 300 -> 285 and 200 -> 190.

Read together, a delegator's position today is: reward 0, downside up to the 5% maximum slash. A reader who lands on the examples page from search sees the opposite picture.

Suggestions:

1. Put the same "rewards paused" notice on Reward Examples, and default the calculator to y = 0, keeping 10% as a labelled historical scenario.

2. Add an example for the current state: stake 500, y = 0, one 5% slash = -25. That makes today's risk/reward explicit.

3. OP-PIP-103 cites the Pythnet sunset planned for Q3 2026 but gives no path for resuming rewards. A line on that in the overview would help.

Question for the team: with rewards at 0 and withdrawals open, what is expected to keep delegated stake in publisher pools, and will OIS restart after the Pythnet sunset or be replaced?

Sources (reviewed 7 Oct 2026):

https://docs.pyth.network/oracle-integrity-staking

https://docs.pyth.network/oracle-integrity-staking/mathematical-representation

https://docs.pyth.network/oracle-integrity-staking/reward-examples

https://forum.pyth.network/t/passed-op-pip-103-pausing-ois-rewards/2471

AI-assisted documentation review; no staking transaction performed.

$PYTH documentation feedback (AI-assisted research): Pyth's 'Why Update Prices' guide describes paying a small protocol fee for every update, while 'Current Fees' says the Core update fee is now zero across mainnet EVM chains as part of the Core sunset.

Could the explainer link that step to Current Fees and distinguish protocol fees from network gas and any data-service charges? Keeping getUpdateFee in the example makes sense; a note that it currently returns zero would clarify the migration without implying every Pyth product is free.

This is a public-documentation review, not a deployed integration or security finding. No affiliation with Pyth. Prepared for the TokenGems feedback contest.

Sources:

https://docs.pyth.network/price-feeds/core/why-update-prices

https://docs.pyth.network/price-feeds/core/current-fees

The pivot in $PYTH from subsidizing to subscriptions is coming in, but the scale is not yet there.

Pyth Pro is booking roughly over $500K a month, and 33% of the DAO treasury goes to monthly PYTH buybacks. So far the Reserve has bought tens of million's of PYTH so far. Pyth has said it targets $500M ARR, and Kalshi and Polymarket using Pyth Pro for settlement is a good sign.

Well, what is the path from ~$4M annualized to buybacks that can turn the float over, and who comes after prediction markets?

People focus on memes but the real Solana play is infrastructure. $PYTH feeding real-world data on-chain, $RENDER R powering decentralised GPU compute, $HNT T building a decentralised wireless network. This ecosystem has layers.

The Pyth Network $PYTH addresses a fundamental friction in high-speed on-chain markets. Its pull-based oracle architecture delivers real-time data feeds directly from first-party sources. This design reduces latency and improves reliability crucial for sophisticated DeFi applications.

The broader market often undervalues this precise infrastructural distinction. Many overlook how $PYTH's approach enables a new class of financial primitives by providing data certainty where older oracle models fall short.