Hyperliquid is a high-performance perps and spot trading L1 with two layers developers actually query: HyperEVM, its EVM-compatible smart contract execution layer, and HyperCore, the native order-book and matching engine. The chain is not slowing down — it currently holds roughly about $7 billion in TVL (DefiLlama, October 2026) and regularly settles billions of dollars of perps volume a day.
Quick verdict: most cost-effective Hyperliquid RPC providers in 2026
Disclosure: Chainstack publishes this page and is one of the six providers listed. Chainstack is best for predictability, and the reasons are in the pricing figures below, not in a benchmark. Dwellir has a lower list price per million standard responses than Chainstack on comparable shared tiers, and that is stated in the comparison. Last reviewed October 29, 2026 (see the changelog at the end).
- Prototypes and early-stage teams (best for predictability: Chainstack): a permanent 3M RU/month free plan with no card, 25 RPS, and flat 1 RU billing per call, so the bill is easy to model before real usage data exists.
- Larger free allowance (Dwellir): a permanent free plan of 100,000 responses per day and one credit per response on shared plans, with HyperCore data billed as add-ons.
- HyperCore data from day one on a free tier (HypeRPC): the free Starter tier (10K requests/month, 5 RPS) covers both layers, using the pre-acquisition rate card that needs re-verifying.
- Free HyperEVM reads with failover (RouteMesh): a free community endpoint for reads; HyperEVM only and no formal SLA.
- All seven HyperCore streams on one endpoint (Quicknode): the broadest HyperCore coverage here, at 20 credits per /info call and with a 30-day trial instead of a permanent free plan.
- HyperEVM-only workloads on an existing stack (Alchemy): permanent free tier, method-weighted compute units (about 10 CU for eth_blockNumber, about 60 CU for eth_getLogs) and no HyperCore access.
- Dedicated infrastructure: Chainstack Dedicated Nodes run $1,489–$2,037/month on mainnet; the Dwellir Dedicated Node Cluster starts at $4,000/month and bundles gRPC and the Order Book API.
2026 is the year the Hyperliquid RPC market actually became a market: Quicknode shipped a HyperCore public beta in January 2026 exposing all seven HyperCore data streams over gRPC, Streams, JSON-RPC, and WebSocket, and multiple aggregators now route Hyperliquid traffic behind a single API. The catch: pricing models diverge sharply enough that the sticker price rarely matches the real bill — method-weighted credit and compute-unit billing can push the effective cost of archive reads or high-frequency polling several multiples above a flat-rate alternative.
This guide compares the six providers worth evaluating for Hyperliquid RPC in 2026, with the actual cost math laid out per method rather than left to marketing pages.
💡 Already using Chainstack? Jump straight to the Hyperliquid tooling docs or deploy your endpoint in minutes.
Why RPC provider choice matters for Hyperliquid
Hyperliquid’s dual-layer design means a provider that’s strong on HyperEVM isn’t automatically strong on HyperCore, and vice versa. Reading order-book depth, fills, and funding rates from HyperCore is a different API surface than calling a smart contract on HyperEVM — some providers cover both, several cover HyperEVM only, and that split matters more here than on a single-layer chain.
Add in how differently providers price the same call, and the true cost of running a Hyperliquid integration can look nothing like the advertised rate. The criteria below separate a production-grade Hyperliquid endpoint from one that will surprise you on the invoice or under load.
- Cost predictability and pricing model — flat request-unit pricing (Chainstack, Dwellir) versus method-weighted compute units or credits (Alchemy, Quicknode, HypeRPC). A workload heavy in HyperCore market-data polling or historical reads pays very differently depending on which model it lands on.
- HyperEVM vs HyperCore coverage — confirm whether a provider actually exposes HyperCore data (order book, fills, funding) or only HyperEVM’s EVM-compatible execution layer. HyperCore access is the single most consequential Hyperliquid-specific gap between providers.
- Throughput stability under load — RPS ceilings and how they hold up during volatility spikes, when trading bots and market makers poll hardest.
- Archive and historical data access — depth and per-request cost of historical HyperEVM state reads, needed for analytics, backfills, and reconciliation.
- Free-tier reality — a permanent standalone plan (Chainstack, Alchemy, Dwellir) versus a time-limited trial that deactivates afterward (Quicknode’s is 30 days). A permanent plan matters for prototyping before committing to a subscription.
- Support quality and responsiveness — a community channel with engineers actually answering on the free tier versus docs-only, and what response-time commitment paid tiers add.
- SLA and compliance — published uptime guarantee, SOC 2 Type II and ISO 27001 for teams that need audited attestation, not self-reported claims.
The table below summarizes public positioning as of October 2026.
| Provider | Best for | Pricing model | Free tier | Dedicated nodes | HyperEVM / HyperCore | SOC 2 / ISO 27001 |
|---|---|---|---|---|---|---|
| Chainstack | Predictable flat-rate billing from the first call; prototypes and regulated buyers | Flat 1 RU/call (2 RU archive and debug/trace) | Permanent — 3M RU/mo, 25 RPS, no card required | Yes, from $810/mo (testnet); mainnet from $1,489/mo | HyperEVM + 46 HyperCore /info query methods (market-data methods such as l2Book via the public API) | Yes / Yes |
| RouteMesh | Free HyperEVM reads and failover across upstream providers | Per-request, by key/chain/method | Free community endpoint (reads) | No (routing layer) | HyperEVM | Not published |
| Dwellir | A larger free allowance (100k responses/day) with HyperCore as paid add-ons | 1 credit per response; shared plans $49–$999/mo; Unlimited Nodes from $119/mo | Permanent — 20 responses/s, 100k responses/day | Yes — Dedicated Node Cluster from $4,000/mo (gRPC + Order Book included) | Both (HyperCore gRPC streaming and Order Book API as add-ons) | Not published |
| HypeRPC | HyperCore data from day one on a free tier (pre-acquisition rate card) | Compute units — $0.50/M CU (EU), $0.75/M CU (JP) (pre-acquisition rate card) | Starter — 10K requests/mo, 5 RPS (pre-acquisition) | Yes | Both | Not published |
| Alchemy | HyperEVM-only workloads for teams already on Alchemy | Compute units (method-weighted) | Permanent | No standard dedicated tier | HyperEVM only | Yes / Not published |
| Quicknode | All seven HyperCore streams from one endpoint, accepting credit billing | Credits (20 credits/info-endpoint call) | 30-day trial only | Yes — Hybrid Dedicated | Both, all 7 HyperCore streams | Yes / Yes |
Key takeaways from the comparison:
- Chainstack offers a permanent free plan with flat, predictable pricing across HyperEVM and its supported HyperCore /info queries, and full SOC 2 Type II and ISO 27001 certification
- RouteMesh routes across multiple upstream providers rather than running its own nodes — useful for teams that want automatic failover without managing several provider relationships
- HypeRPC is the only Hyperliquid-exclusive specialist here, covering both HyperEVM and HyperCore with dedicated infrastructure
- Alchemy covers HyperEVM only — no HyperCore data — but has the most mature general-purpose EVM tooling
- Quicknode covers all seven HyperCore streams from one unified endpoint but has no permanent free tier and credit-weighted billing
Hyperliquid RPC rate limits and free tiers compared
Rate limits and billing units differ between providers, so the same number of calls can cost different amounts. This table repeats the figures from the provider sections below; no new prices are introduced.
| Provider | Free tier | Paid shared plans (as listed in this article) | Billing unit |
|---|---|---|---|
| Chainstack | Permanent: 3M RU/mo, 25 RPS, no card | Growth $49/mo (20M RU, 250 RPS); Pro $199/mo (80M RU, 400 RPS); Business $499/mo (200M RU, 600 RPS); Enterprise from $990/mo | Flat 1 RU per call; 2 RU for archive and debug/trace |
| RouteMesh | Free community endpoint (read requests) | Premium tier priced per request by key type, chain and method | Per request |
| Dwellir | Permanent: 20 responses/s, 100,000 responses/day | Developer $49/mo (25M responses, 100/s); Growth $299/mo (150M, 500/s); Scale $999/mo (500M, 5,000/s); Unlimited Nodes from $119/mo | 1 credit per response; HyperCore add-ons billed on top |
| HypeRPC | Starter: 10K requests/mo, 5 RPS (pre-acquisition) | Professional $99/mo (1M requests, 50 RPS); overage $0.50 per M CU (EU), $0.75 per M CU (JP) | Compute units |
| Alchemy | Permanent free tier | Not itemized in this article | Compute units: about 10 CU (eth_blockNumber) to about 60 CU (eth_getLogs) |
| Quicknode | 30-day trial only | $49/mo up to $999/mo (highest non-custom tier) | Credits: 20 per /info call; gRPC streaming 10 credits per 0.1 MB |
List price per million standard calls (derived). For the two flat-billing providers, dividing the monthly price by the included units gives a comparable figure for standard calls, before overage and excluding archive calls (2 RU on Chainstack) and Dwellir HyperCore add-ons. Alchemy, Quicknode and HypeRPC bill by method weight, and this article lacks the included-unit figures needed to derive a like-for-like number.
| Plan | Monthly price | Included units | List price per million standard calls |
|---|---|---|---|
| Chainstack Growth | $49 | 20M RU | $2.45 |
| Dwellir Developer | $49 | 25M responses | $1.96 |
| Chainstack Pro | $199 | 80M RU | $2.49 |
| Dwellir Growth | $299 | 150M responses | $1.99 |
| Chainstack Business | $499 | 200M RU | $2.50 |
| Dwellir Scale | $999 | 500M responses | $2.00 |
Dwellir comes out about 20% lower per million standard calls on these shared tiers. Chainstack’s case rests on predictability across methods and on a free plan that includes HyperEVM and the supported HyperCore /info queries, while Dwellir bills HyperCore streaming and Order Book access separately.
How we compared Hyperliquid RPC providers on cost
Disclosure: this comparison is published by Chainstack. Chainstack is one of the six providers listed and is labelled best for predictability, so read the picks as a vendor perspective and check the figures against the official pricing pages linked in each provider section.
How the picks were made. The table has no rank column because the picks depend on the workload. Each pick comes from the use-case sections of this article (cost-conscious teams, trading bots, regulated deployments) and from the pricing and free-tier figures above. The picks are editorial judgment, not a benchmark result and not a numeric ranking.
Criteria. Each provider is compared on pricing model, free tier, published rate limits (RPS), HyperEVM and HyperCore coverage, dedicated node pricing, support, SLA and security certifications.
Sources. Prices, limits and certification statements come from each provider’s public pricing, documentation and status pages, as summarized in the provider sections. The summary table reflects public positioning as of October 2026. HypeRPC figures are the pre-acquisition rate card. Plans change often, so confirm current prices on the official pricing pages before you buy.
Not covered. Other providers may also support Hyperliquid. dRPC appears in the October 2026 latency report cited below but is not one of the six providers compared here.
Latency data. Cost-effectiveness in this article is judged on the pricing figures above. The latency report is context only, covers three of the six providers, and does not measure cost. See the latency context section.
How to choose a Hyperliquid RPC provider
1. Pricing model — flat vs weighted, and what that means at scale
Chainstack bills every Hyperliquid call at a flat 1 RU (2 RU for archive and debug/trace calls). Alchemy, Quicknode, and HypeRPC weight by method — a HyperCore info-endpoint call costs Quicknode 20 credits, while a heavier HyperEVM read like eth_getLogs costs roughly 60 compute units on Alchemy versus roughly 10 for a simple eth_blockNumber call. The formula for comparing them: true cost per call = (method’s CU/credit weight × price per unit) ÷ total units in your plan — multiply by monthly call volume for the real bill. Run your own numbers through the Chainstack cost calculator rather than trusting a headline rate.
| Provider | Unit weight (info/read call) | Effective shape at volume |
|---|---|---|
| Chainstack | 1 RU flat | Predictable — every call costs the same regardless of method |
| HypeRPC | Variable CU by call complexity | $0.50–$0.75 per million CU overage, region-dependent |
| Quicknode | 20 credits (info endpoint) | Heavier for market-data polling than a flat-rate model |
| Alchemy | ~10–60 CU depending on method | Archive/log-heavy calls cost multiples of a simple read |
2. Latency consistency vs. average latency
For trading bots and market makers, tail latency matters more than the advertised average. A provider that’s fast most of the time but spikes during volatility — exactly when Hyperliquid sees its heaviest order flow — will cost more in missed fills than a slightly slower but consistent one. Check compare.chainstack.com for a live, continuously-updated ranking across tracked chains including Hyperliquid, and benchmark p95/p99 response times against your own target region before committing. The latency context section below shows what the October 2026 benchmark did and did not measure.
3. Archive and historical data access
Analytics, reconciliation, and backfill workloads live on archive access. Confirm depth (how far back), method coverage, and per-request cost before assuming a provider’s archive tier covers your use case — Chainstack prices archive reads at 2 RU versus 1 RU for standard calls, a fixed and known multiplier rather than a variable one.
4. Free-tier reality
Chainstack and Alchemy run permanent free plans with no card required, and Dwellir also has a free plan (100,000 responses per day). HypeRPC’s Starter tier (10K requests/month, 5 RPS) is also standalone. Quicknode’s entry point is explicitly a 30-day trial, not a permanent tier — fine for a quick prototype, less so for an evaluation that runs longer than a month.
5. When dedicated infrastructure matters
Shared endpoints stop being enough once you hit sustained load above the free-tier RPS ceiling, need isolated infrastructure for compliance reasons, or require guaranteed throughput during volatility spikes when every other bot on a shared endpoint is polling just as hard. Dedicated Nodes for Hyperliquid start from $810/month (testnet), scaling to $1,489/month for mainnet archive and $2,037/month for mainnet full nodes; HypeRPC and Quicknode’s Hybrid Dedicated also offer isolated tiers.
6. HyperEVM vs HyperCore coverage
HyperEVM versus HyperCore coverage is the criterion unique to Hyperliquid. If your workload only touches smart contracts, HyperEVM-only coverage (Alchemy) is fine. If you need order-book depth, fills, or funding data, you need a provider that exposes HyperCore specifically. HypeRPC, Quicknode, and Dwellir (through paid add-ons) serve HyperCore market data and streams. Chainstack serves HyperEVM plus 46 HyperCore /info query methods (account and state queries such as clearinghouseState and openOrders), while proprietary market-data methods such as l2Book and recentTrades are only available from Hyperliquid’s public API. RouteMesh currently routes HyperEVM traffic only.
7. Support quality — tiered, not “SLA yes/no”
Chainstack runs community support (Discord and Telegram with engineers active in-channel) on the free Developer plan, standard email response from Growth ($49/mo), and 24/7 Enterprise support with fast SLA-backed response on Enterprise. Verify each competitor’s current tier names and response commitments directly — support-tier structures change more often than headline pricing.
Latency context for cost decisions
Scope and disclosure. This section is context, not a cost ranking. Chainstack operates the monitoring service, the public dashboard and the scoring formula behind the figures below, and Chainstack is one of the four providers measured. The report is Hyperliquid RPC provider benchmark: latency and uptime, October 2026 (Research Note No. 051) by the Chainstack Labs Research Team. It measured Chainstack, Alchemy, Quicknode and dRPC. Of the six providers in this article, only Chainstack, Alchemy and Quicknode were measured; RouteMesh, Dwellir and HypeRPC were not.
Method in brief. Probes ran every 1 to 3 minutes for 7 days ending October 8, 2026, from three cloud regions (Germany, the US and Japan), against each provider’s first paid plan. A request counts as failed after 55 seconds. The score is response time divided by the cube of the success rate per region (lower is better), averaged across the three regions. The collection code is open source (compare-dashboard-functions).
Probe success rate is not platform uptime. The probe success rate is the share of the probe requests that succeeded in the 7-day window. Platform uptime is a different measure: status.chainstack.com shows the Hyperliquid Mainnet component at 99.99%+ over the 90 days to October 29, 2026 (99.99%+ platform uptime, as displayed on the status page, which covers Chainstack only). Status-page figures are for the Mainnet component of each chain; testnet and devnet components are tracked separately. Neither figure is a contractual SLA or a guarantee, and this article states no uptime figures for the other providers.
| Provider | Score (lower is better) | Probe success rate | P50 | P95 | P99 | DE P95 | US P95 | JP P95 |
|---|---|---|---|---|---|---|---|---|
| Chainstack | 0.019 | 99.74% | 10 ms | 22 ms | 82 ms | 18 ms | 31 ms | 18 ms |
| Alchemy | 0.036 | 100.00% | 19 ms | 40 ms | 131 ms | 28 ms | 43 ms | 51 ms |
| Quicknode | 0.065 | 99.98% | 53 ms | 76 ms | 122 ms | 19 ms | 81 ms | 128 ms |
| dRPC (measured, not one of the six here) | 0.124 | 100.00% | 23 ms | 129 ms | 248 ms | 76 ms | 127 ms | 183 ms |
- Chainstack recorded the lowest latency at every percentile and in every region, and Alchemy was second on score (1.9 times Chainstack’s score).
- Chainstack also recorded the lowest probe success rate in the set at 99.74%, against 100.00% for Alchemy and dRPC and 99.98% for Quicknode (the report and the Compare dashboard label this column availability). The report estimates this costs Chainstack about 0.8% of its score and does not change the rank. For workloads where every failed request matters, that gap is worth weighing.
- Japan was the weakest region for all three competitors: Alchemy 51 ms, Quicknode 128 ms and dRPC 183 ms at P95.
What this benchmark does not show. It does not measure cost, throughput under load or correctness. It covers entry paid plans only, uses synthetic probes from cloud regions rather than end-user latency, and uses a single 7-day window. Chainstack chose the scoring formula, so a different formula could reorder close pairs. It says nothing about RouteMesh, Dwellir or HypeRPC. Use compare.chainstack.com for current figures and how to measure RPC latency to test from your own region. Read the full report for the method and raw tables.
Choose the right Hyperliquid RPC provider by use case
For cost-conscious teams and early-stage builders
Early-stage teams need a free tier that doesn’t expire and a pricing model that stays predictable as usage grows from zero to production. A time-limited trial forces a pricing decision before you have real usage data to base it on.
Chainstack’s permanent Developer plan (3M RU/month, no card) and flat per-call billing make it, in our assessment, the easiest to model early and a plan that scales without a pricing-model migration later. The comparison above shows that Dwellir’s list price per million standard responses is lower than Chainstack’s on comparable shared tiers, so the flat-rate advantage here is predictability across methods, not the lowest unit price. HypeRPC’s free Starter tier is a reasonable second option if you specifically need HyperCore data from day one.
For trading bots and market makers
Trading infrastructure on Hyperliquid needs both layers — HyperCore for order-book state and fills, HyperEVM for any on-chain settlement logic — with throughput that holds up specifically during the volatility spikes when order flow is heaviest and every competing bot is polling the same endpoint just as hard.
HypeRPC, Quicknode, and Dwellir (gRPC streaming and Order Book API add-ons) expose HyperCore market data alongside HyperEVM, while Chainstack covers HyperEVM and the HyperCore account and state queries. Quicknode’s all-seven-stream gRPC coverage is the broadest from a single unified endpoint; HypeRPC is purpose-built for exactly this workload; Chainstack’s flat pricing keeps the bill predictable even as polling volume climbs during high-volatility periods.
For regulated and enterprise deployments
High-frequency trading and market-making desks that touch Hyperliquid at institutional volume carry compliance obligations most of the providers in this comparison don’t publicly address. Audited certifications, not self-reported security claims, are what a vendor-risk review actually asks for.
Chainstack and Quicknode are the only two providers here with both SOC 2 Type II and ISO 27001 published. HypeRPC, RouteMesh, and Dwellir do not publish either — a real gap worth naming honestly rather than assuming coverage that isn’t documented.
Provider-by-provider breakdown
Chainstack: best for predictable flat-rate billing
Verdict: the most predictable bill from the first call (flat 1 RU per call, permanent 3M RU/month free plan); the trade-offs are a higher Dedicated Node entry price on Hyperliquid and HyperCore coverage limited to /info account and state methods. Chainstack publishes this page, so check the official Chainstack pricing page yourself.
Chainstack runs Hyperliquid RPC infrastructure covering HyperEVM and 46 HyperCore /info query methods (proprietary market-data methods such as l2Book stay on Hyperliquid’s public API), billed at a flat 1 RU per call (2 RU for archive and debug/trace calls) regardless of method — the same predictable model across all 70+ chains Chainstack supports, not a Hyperliquid-specific exception.
The free Developer plan includes 3M RU/month at 25 RPS with no card required, scaling through Growth ($49/mo, 20M RU, 250 RPS), Pro ($199/mo, 80M RU, 400 RPS), Business ($499/mo, 200M RU, 600 RPS), up to Enterprise (from $990/mo). Dedicated Nodes for Hyperliquid start from $810/month for isolated, unmetered access on testnet, scaling to $1,489–$2,037/month for mainnet. Chainstack also holds SOC 2 Type II and ISO 27001 certification, and a Hyperliquid testnet faucet is available for development.
Platform uptime: 99.99%+ platform uptime (status.chainstack.com, Hyperliquid Mainnet, 99.99%+ over the 90 days to October 29, 2026). This is the status-page figure, separate from the benchmark probe success rate in the latency section and from contractual SLAs such as the Enterprise SLA.
Chainstack limitations: Dedicated Nodes for Hyperliquid sit at a higher entry price ($810/mo testnet, $1,489+/mo mainnet) than the generic Dedicated Nodes baseline on other chains, reflecting the chain’s own infrastructure requirements — worth budgeting for if you know you’ll need isolation early.
Chainstack fit by workload:
- Cost-conscious teams: Excellent — a permanent free tier with flat billing across HyperEVM and the supported HyperCore info queries.
- Trading bots: Strong — predictable costs even as polling volume spikes during volatility, though Dedicated Nodes are worth evaluating for latency-critical desks.
- Regulated deployments: Excellent — dual certification and a contractual 99.9% quarterly SLA on the Enterprise plan give compliance teams something to actually cite.
RouteMesh: best for free HyperEVM reads with failover
Verdict: a free community endpoint for HyperEVM reads and automatic failover across upstream operators; the trade-offs are HyperEVM only and no formal SLA. Official RouteMesh site and pricing.
RouteMesh is an RPC aggregation and routing layer rather than a node operator — it routes requests across 10-14 upstream providers per method, including Chainstack, with real-time health scoring and automatic failover. RouteMesh confirms full support for Hyperliquid’s HyperEVM (chain ID 999) among its 1000+ supported chains, with a free community endpoint at lb.routeme.sh/rpc/evm/999.
Pricing is split between the free community endpoint (read requests) and a premium tier priced per request by key type, chain, and method. For teams that want automatic failover across many upstream operators without managing separate provider relationships, that’s the pitch.
RouteMesh limitations: HyperEVM only — no HyperCore market-data routing currently. No formal SLA; archive depth depends on whichever upstream serves a given request rather than being something you configure directly.
RouteMesh fit by workload:
- Cost-conscious teams: Good — a free community endpoint covers basic HyperEVM reads.
- Trading bots: Moderate — no HyperCore coverage limits it to the EVM side of a trading stack.
- Regulated deployments: Limited — no published compliance certifications.
Dwellir: best for a larger free allowance and lower shared-plan unit price
Verdict: a permanent free plan (100,000 responses per day) and a lower list price per million standard responses than Chainstack on comparable shared tiers; the trade-offs are HyperCore billed as add-ons and a $4,000/mo dedicated cluster entry point. Official Dwellir pricing.
Dwellir is a Swedish RPC provider with a long Polkadot track record that now serves 150+ networks, Hyperliquid among them. On HyperEVM (chain ID 999) Dwellir serves HTTPS and WebSocket with the debug and trace namespaces enabled on every API plan. For HyperCore it sells separate products: a gRPC streaming API with raw node feeds and replay (from $299/mo plus usage), an Order Book API for L2, L4, and BBO data (from $199/mo plus usage), and archive node files.
Shared plans bill one credit per response regardless of method: Free covers 20 responses/second and 100,000 responses per day, Developer is $49/mo (25M responses, 100/second), Growth $299/mo (150M, 500/second), and Scale $999/mo (500M, 5,000/second). Unlimited Nodes start at $119/mo flat. A Dedicated Node Cluster for Hyperliquid starts at $4,000/mo, unmetered, with gRPC and the Order Book API included and no usage fees.
Dwellir limitations: HyperCore data is an add-on billed on top of the RPC plan, so the real monthly cost depends on how many streams you run. The $4,000/mo cluster entry point is above Chainstack’s mainnet Dedicated Nodes ($1,489–$2,037/mo), although it bundles gRPC and Order Book access. We found no SOC 2 or ISO 27001 attestation on its public pricing and dedicated-cluster pages.
Dwellir fit by workload:
- Cost-conscious teams: Good — a permanent free plan and one-credit-per-response billing, but HyperCore add-ons raise the bill once you need streaming or order book data.
- Trading bots: Strong — gRPC streaming with replay and an Order Book API from Singapore and Tokyo cover the HyperCore side directly.
- Regulated deployments: Limited — Dwellir’s marketing pages state a 99.99% uptime SLA (contract terms not read), but no audited certification is published.
HypeRPC: best for HyperCore from a Hyperliquid-only specialist
Verdict: both HyperEVM and HyperCore from a free Starter tier on a Hyperliquid-only specialist; the trade-offs are pre-acquisition pricing that needs re-verifying and the smallest free tier here. Official HypeRPC site and pricing.
HypeRPC, built by Imperator, describes itself as the first dedicated RPC provider for Hyperliquid — shared and dedicated nodes covering both HyperEVM and HyperCore, direct sentry peering with sub-500-microsecond validator connections, and orderbook WebSocket access up to 100 levels of depth. HypeRPC was acquired by Hyperdash in 2026 — pricing below reflects the pre-acquisition public rate card and should be verified directly before committing, since the acquisition may change plans or branding.
Pricing ran on compute units: a free Starter tier (10K requests/month, 5 RPS) for testing, a Professional tier at $99/month (1M requests, 50 RPS), and CU overage at $0.50 per million in the EU region or $0.75 per million in JP.
HypeRPC limitations: No SOC 2 or ISO 27001 published — a real gap for institutional buyers. Being Hyperliquid-only means no multi-chain fallback if you also need coverage elsewhere.
HypeRPC fit by workload:
- Cost-conscious teams: Good — a real free tier, though it’s the smallest here (10K requests/month).
- Trading bots: Excellent — direct sentry peering and deep orderbook access are built for exactly this workload.
- Regulated deployments: Limited — no audited certifications to point a compliance review at.
Alchemy: best for HyperEVM-only workloads on an existing Alchemy stack
Verdict: a permanent free tier and mature general-purpose tooling for HyperEVM; the trade-offs are no HyperCore access and method-weighted compute-unit billing. Official Alchemy pricing.
Alchemy’s Hyperliquid support is HyperEVM-only — no HyperCore query routing or dedicated HyperCore infrastructure. For teams building EVM-native contracts and DeFi protocols on HyperEVM specifically, Alchemy’s mature general-purpose tooling (per-method analytics, request tracing) is a familiar option if you already run other chains through it.
Pricing runs on compute units, method-weighted — a simple read like eth_blockNumber costs around 10 CU, while eth_getLogs runs closer to 60 CU. The free tier is permanent, a genuine positive next to Quicknode’s trial-only entry point.
Alchemy limitations: No HyperCore access at all rules it out for any workload touching order-book data. No standard dedicated node tier. In our assessment, Alchemy’s per-CU pricing can cost more than a flat-rate plan at meaningful production volume on heavier methods — worth running your own numbers through a cost calculator before committing.
Alchemy fit by workload:
- Cost-conscious teams: Moderate — permanent free tier, but CU costs climb fast on heavier methods.
- Trading bots: Limited — no HyperCore means no order-book or fills data at all.
- Regulated deployments: Moderate — SOC 2 Type II published, ISO 27001 is not.
Quicknode: best for full HyperCore coverage from one endpoint
Verdict: the broadest HyperCore coverage (all seven streams) on one endpoint; the trade-offs are a 30-day trial instead of a permanent free tier and credit-weighted billing. Official Quicknode pricing.
In our assessment, Quicknode is the most feature-complete provider in this comparison for Hyperliquid specifically — its January 2026 HyperCore public beta exposes all seven HyperCore data streams through gRPC, Streams, JSON-RPC, and WebSocket from a single endpoint that also serves HyperEVM. For teams that want every available data surface from one provider, that’s a real advantage.
Billing is credit-based: all supported /info endpoint methods cost 20 credits per request, and gRPC streaming bills at 10 credits per 0.1 MB of filtered data. Paid plans run from $49/month up to $999/month for the highest non-custom tier.
Quicknode limitations: No permanent free tier — the entry point is a 30-day trial, not a standalone plan. Method-weighted credit billing can surprise teams once HyperCore streaming or heavy market-data polling enters the mix — worth modeling before committing to a tier.
Quicknode fit by workload:
- Cost-conscious teams: Limited — no permanent free plan to prototype against.
- Trading bots: Excellent — all seven HyperCore streams over gRPC from the same endpoint that serves HyperEVM.
- Regulated deployments: Strong — both SOC 2 Type II and ISO 27001 published.
Chainstack alternatives for Hyperliquid RPC
If Chainstack is not the right fit, these are the head-to-head differences between Chainstack and each alternative, using the figures in this article:
- Chainstack vs Quicknode: Quicknode covers all seven HyperCore streams, while Chainstack serves HyperEVM plus 46 HyperCore /info query methods. Quicknode bills 20 credits per /info call and offers a 30-day trial; Chainstack bills a flat 1 RU and has a permanent free plan. Both publish SOC 2 Type II and ISO 27001. In the October 2026 report, Chainstack’s P95 was 22 ms against Quicknode’s 76 ms, with Quicknode at a 99.98% probe success rate against Chainstack’s 99.74%.
- Chainstack vs Alchemy: Alchemy is HyperEVM only and bills method-weighted compute units; Chainstack adds the supported HyperCore /info queries at a flat rate. Both have permanent free tiers. Alchemy publishes SOC 2 Type II but not ISO 27001, and has no standard dedicated tier. In the report, Alchemy recorded a 100.00% probe success rate and a 40 ms P95, against Chainstack’s 99.74% and 22 ms.
- Chainstack vs Dwellir: Dwellir has a lower list price per million standard responses on shared plans, and a 100,000 responses/day free plan, but bills HyperCore gRPC (from $299/mo) and the Order Book API (from $199/mo) on top. Its dedicated cluster starts at $4,000/mo against Chainstack’s $1,489–$2,037/mo mainnet Dedicated Nodes, although the cluster bundles gRPC and Order Book access. Dwellir publishes no SOC 2 or ISO 27001 attestation. Dwellir was not part of the latency report.
- Chainstack vs HypeRPC: HypeRPC is Hyperliquid-only with HyperCore coverage, sentry peering and orderbook WebSocket depth; Chainstack runs 70+ chains and a flat 1 RU billing model. HypeRPC publishes no SOC 2 or ISO 27001, and its pricing is the pre-acquisition rate card. HypeRPC was not part of the latency report.
- Chainstack vs RouteMesh: RouteMesh is a routing layer that sends requests across upstream providers (Chainstack among them) and covers HyperEVM only, with no formal SLA. Chainstack runs its own Hyperliquid infrastructure. RouteMesh was not part of the latency report.
Getting started with Hyperliquid on Chainstack
Deploy a production Hyperliquid endpoint in a few steps and build better with Hyperliquid on Chainstack:
- Log in to the Chainstack console (or create an account).
- Create a new project
- Select Hyperliquid as your blockchain protocol
- Choose network: Hyperliquid Mainnet or Testnet
- Deploy the node
- Open Access and credentials and copy your HTTPS and WebSocket endpoints
For latency-critical trading workloads, evaluate Dedicated Nodes after the initial deploy — available inside the same project.
import { InfoClient, HttpTransport } from "@nktkas/hyperliquid";
const transport = new HttpTransport({ url: "YOUR_CHAINSTACK_ENDPOINT" });
const info = new InfoClient({ transport });
// Fetch perp market metadata from HyperCore (an /info method Chainstack serves)
const meta = await info.meta();
console.log(meta);
📖 For the full integration guide, see the Chainstack Hyperliquid tooling documentation.
Need testnet funds? Grab some from the Chainstack Hyperliquid faucet.
🤖 You can also access Chainstack Hyperliquid RPC directly from Claude, Cursor, Codex, Windsurf, Gemini CLI, GitHub Copilot, Antigravity, Claude.ai, or ChatGPT using Chainstack MCP. For a fuller agent stack, see the Chainstack Agents page.
Conclusion
The Hyperliquid RPC market split cleanly in 2026 along two axes: whether a provider covers HyperCore alongside HyperEVM, and whether its pricing stays flat as usage grows or weights by method. Those two questions, more than brand recognition, decide which provider actually fits a given workload.
- Cost-conscious teams and prototypes → Chainstack’s permanent free tier and flat billing
- Trading bots and market makers needing full HyperCore depth → Quicknode, Dwellir, or HypeRPC
- Regulated or institutional deployments → Chainstack or Quicknode, the only two with both SOC 2 Type II and ISO 27001 published
- HyperEVM-only DeFi and contract workloads → Alchemy, if already standardized on it for other chains
FAQ
Q: What’s the difference between HyperEVM and HyperCore RPC access?
HyperEVM is Hyperliquid’s EVM-compatible smart contract layer — standard eth_* JSON-RPC methods apply. HyperCore is the native order-book and matching engine, queried through a separate /info API for market data, fills, and funding rates. Some providers (Alchemy) cover HyperEVM only; others (HypeRPC, Quicknode, Dwellir) cover both, and Chainstack covers HyperEVM plus a subset of HyperCore /info queries.
Q: Which Hyperliquid RPC provider has the best free tier?
Chainstack’s Developer plan (3M RU/month, 25 RPS, no card required) and Dwellir’s Free plan (100,000 responses per day, about 3M a month) are the largest permanent free tiers here; Chainstack’s covers HyperEVM and the supported HyperCore info queries, while Dwellir sells HyperCore streaming as paid add-ons. HypeRPC’s free Starter tier was smaller (10K requests/month, pre-acquisition) but Hyperliquid-specific. Quicknode’s entry point is a 30-day trial, not a permanent free plan.
Q: How do I migrate from a public Hyperliquid RPC endpoint to a managed provider?
Deploy a node with your chosen provider, copy the new HTTPS/WebSocket endpoint, and swap it into your existing SDK configuration — the API surface is the same. Test against testnet first using the Chainstack Hyperliquid faucet before cutting over production traffic.
Q: What latency should I expect from a Hyperliquid RPC provider?
Latency varies by region and provider, and averages hide the tail latency that matters most for trading. Check compare.chainstack.com for live rankings, and benchmark p95/p99 response times from your own target region rather than relying on a single average figure. The latency context section summarizes the October 2026 report, which measured Chainstack, Alchemy, Quicknode and dRPC.
Q: Do Hyperliquid RPC providers publish security certifications?
Chainstack and Quicknode both publish SOC 2 Type II and ISO 27001. Alchemy publishes SOC 2 Type II but not ISO 27001. HypeRPC, RouteMesh, and Dwellir do not publish either certification.
Q: Is flat-rate or credit-based pricing cheaper for Hyperliquid RPC?
The answer depends on your method mix. Flat-rate billing (Chainstack, Dwellir) is more predictable and can be cheaper for archive-heavy or high-volume polling workloads, since method-weighted models (Alchemy, Quicknode, HypeRPC) charge multiples of a base rate for heavier calls. Run your actual call mix through the Chainstack cost calculator to see the real number for your workload.
Additional resources
- Hyperliquid tooling documentation — Chainstack Docs
- Hyperliquid available methods — Chainstack Docs
- Hyperliquid API reference overview — Chainstack Docs
- Hyperliquid official documentation
- More Hyperliquid articles on the Chainstack Blog
- Chainstack pricing
Changelog
We review this comparison regularly so that the information stays accurate, clear, and easy to verify. This page was reviewed to improve how the comparison is structured and sourced.
- October 29, 2026: Added a quick verdict by cost scenario, a Best for column in the comparison table, a rate limits and free tiers table with derived list prices per million standard calls, a how we compared section, a latency context section summarizing the October 2026 Hyperliquid benchmark report (probe success rate shown separately from Chainstack platform uptime from status.chainstack.com), a Chainstack alternatives section, per-provider verdicts with official pricing links, and this changelog. Corrected vendor pricing, limits and certification details, and reworded several sentences for clarity and to separate editorial judgment from stated facts.
- March 4, 2026: Original publication.