The 9 best Solana RPC providers in 2026: a decision guide

TL;DR: Which Solana RPC provider should you pick?

What production Solana teams need from an RPC provider:

• Fast, reliable reads. Account reads form the largest share of typical Solana traffic, so read latency, rate limits, and history depth decide how your app feels.
• Streaming instead of polling. Solana's own indexing docs put Yellowstone gRPC at sub-100ms versus roughly 200-400ms for WebSocket subscriptions, which is why trading, MEV, and indexing teams stream.
• More than one chain. 1 in 3 crypto developers now works on multiple chains, up from under 10% in 2015, and Solana was the #1 ecosystem for new developers in 2024 with 83% year-over-year growth. Many Solana teams read from other chains too.

Which Solana RPC provider should you pick?

• Best overall: Alchemy. 99.99% uptime, multi-region infrastructure with autonomous failover, 100% of transactions sent through staked connections, and the lowest average latency globally, in US East, and in US West on the live Solana benchmarks (October 15, 2026 snapshot, five providers). Powers Phantom, Solflare, Robinhood, OpenSea, and Circle.
• For production reads and history: use Alchemy RPC. Up to 10x faster on heavy reads like getProgramAccounts and up to 20x faster on archive calls.
• For account-state history and backfills: use the Solana Account Archive, which covers account state that transaction-history archives don't.
• For streaming: use Alchemy Solana gRPC, Yellowstone-compatible at $75/TB with no monthly minimum, the lowest per-TB or per-GB rate among the providers we compared that publish one (Triton lists $0.08/GB, about $80/TB). Flat-fee plans can cost less at high volume, so check the gRPC pricing table below.
• For multi-chain teams: use one Alchemy account for Solana plus 100+ chains, with one bill.
• To try it first: the free tier includes 30M CU/month.

Where another provider can be a reasonable second choice:

• Helius: for Solana-only teams that want its enhanced NFT and token parsing APIs and direct leader connections. Alchemy covers the same core RPC, history, and gRPC surface and adds other chains.
• Triton: for HFT and MEV teams that want bare-metal infrastructure and shred streaming. It is built for landing and streaming workloads rather than general-purpose reads.
• QuickNode: for multi-chain teams that value breadth (80+ chains) and built-in analytics. Alchemy leads it on average latency in our October 15, 2026 benchmark.
• Chainstack: for teams that need self-hosted or dedicated deployment options, or a low-cost entry to archive access on its $49/month Growth plan, with Yellowstone gRPC as an add-on.
• Ankr: for cost-sensitive teams reading recent state only, since its Solana nodes keep roughly 16 hours of transaction history.
• Syndica: for Solana-only enterprise teams that are comfortable with a sales-led motion.
• dRPC: as a redundancy layer in front of a primary provider, not as the primary.
• Blockdaemon: for institutions whose RPC needs sit alongside custody, staking, or MPC requirements.

Looking for a Helius alternative? Alchemy covers the same core surface (RPC, genesis-depth history, Yellowstone-compatible gRPC, webhooks) and adds account-state history and 100+ other chains. We also compare Helius, QuickNode, Triton, Ankr, Syndica, Chainstack, dRPC, and Blockdaemon below, with dated sources for every price and benchmark.

If you only read one section, read the How to choose checklist below.

Why your choice of Solana RPC provider matters

Solana now processes around 170 million non-vote transactions a day (5.2 billion in August 2026, a monthly record), and the full transaction ledger has grown past 400TB. The network keeps raising its own ceiling, too. The SIMD-0286 upgrade lifted the per-block compute limit from 60M to 100M compute units in July 2026, and Alpenglow, the consensus upgrade targeting roughly 150ms finality, is flagged by Solana as a breaking change for indexers when it activates.

At that volume and pace, a wrong RPC choice surfaces as dropped transactions during spikes, stale reads, missing history, and surprise bills when traffic multiplies. Public endpoints are fine for prototypes. Production apps need a private endpoint with an SLA, predictable pricing, and archival depth, and that's the choice this guide is built around.

At-a-glance comparison

Pricing, SLAs, and limits were checked against each provider's public pricing and docs pages in September and October 2026. They change frequently; verify before committing.

Solana RPC latency by region

The most useful latency number is the one measured from your users' region, so here is the same benchmark broken out by region. These figures come from our public Solana benchmarks, a rolling 24-hour window. This snapshot covers the 24 hours ending October 15, 2026 at 17:40 UTC. Each cell shows average latency in milliseconds, with P95 in parentheses.

What the numbers say:

• Alchemy had the lowest average latency globally (8.07 ms), in US East (7.40 ms), and in US West (10.76 ms), and the lowest P95 in all three. Success rate was 100% globally.
• In EU Central and AP Southeast, averages across the leading providers sit within a few milliseconds of each other, and the table above lists every provider's result. If your users are concentrated in Europe or Asia-Pacific, run your own traffic against Alchemy's free 30M CU tier to confirm latency from those regions.
• Triton averaged 17.89 ms in US East. Its global average is higher because of EU Central (139.95 ms) and AP Southeast (181.89 ms).

How the benchmark is run: we measure seven read methods (getAccountInfo, getBalance, getBlockHeight, getLatestBlockhash, getMultipleAccounts, getSlot, and getSlotLeader) from AWS in us-east-1, us-west-2, eu-central-1, and ap-southeast-1, using standard paid accounts, warmed connections, no retries, and an 8-second timeout. It refreshes every few minutes, so the live dashboard will differ from this snapshot. It doesn't measure archive calls, getProgramAccounts, gRPC, or transaction landing. For the full methodology, see how we benchmark RPC performance and our Solana read benchmark write-up.

How to choose a Solana RPC provider

Work down this checklist in order. The first place a provider fails is usually the right place to disqualify them.

1. Uptime SLA. Require ≥99.9% with financial penalties, and check whether the number is a contractual SLA or a marketing claim. We commit to 99.99%.
2. Latency from your regions, under load. Test P90 from your actual user regions with traffic at 10x your current peak. Sub-150ms is table stakes; sub-100ms matters for trading and MEV. The live benchmarks publish continuously updated Solana latency and success rates measured from US East, US West, EU Central, and AP Southeast, so you can check dated numbers before running your own test.
3. Archival depth. Decide upfront which kind of history you need. Transaction and block history back to genesis is now documented by several providers. Account-state history (what did this account hold at slot X?) is much rarer, because validators discard old account state. Our guide to Solana archival data covers the difference in depth.
4. Pricing model. CU-based, transparent pricing (for example, our $0.52/1M CU) gives the most predictable bills. Be skeptical of per-request models, and check what's plan-gated, since some providers sell streaming only inside $499/month tiers. Different providers also count credits differently.
5. Websocket and streaming reliability. If you depend on subscriptions, demand over 95% message delivery and reconnect semantics you can rely on. For high-volume real-time data, compare webhooks, websockets, and gRPC before assuming you need a stream at all.
6. Tooling adjacency. Webhooks, request logs, alerts, and enriched data APIs are usually where the real productivity gap lives between providers.
7. Migration friction. Solana uses standard JSON-RPC, so most migrations are just an endpoint swap. Verify any provider-specific extensions you use are supported.

The 9 best Solana RPC providers

1. Alchemy: built for mission-critical Solana and multi-chain reads

We power 71%+ of the top onchain apps, process $1T+ in annual transactions, and serve 100M+ humans across the crypto ecosystem. Phantom, Solflare, Robinhood, OpenSea, and Circle all run on Alchemy. The Solana stack is purpose-built from the ground up, combining 9+ years of infrastructure experience with deep Solana ecosystem expertise from leading engineers at DexterLab and Bware Labs.

Where it shows up in production:

• Up to 10x faster on heavy methods like getProgramAccounts and getTokenAccountsByOwner, and up to 20x faster on archive calls, per our published Solana infrastructure benchmarks
• Lowest average latency globally, in US East, and in US West on our public benchmarks, which test Solana JSON-RPC against Chainstack, Helius, QuickNode, and Triton from four regions (US East, US West, EU Central, AP Southeast) over a trailing 24-hour window. In the October 15, 2026 snapshot we averaged 8.07 ms globally at 100% success. Results in EU Central and AP Southeast were close across providers, as shown in the regional table above.
• 99.99% uptime and 2x+ higher throughput across the platform (source)
• 100% of transactions sent through staked connections for maximum landing rate
• Yellowstone-compatible gRPC with delivery 5-15ms faster on average than alternatives (measured in us-east-1), automatic backfill on reconnect, and pay-as-you-go pricing at $75/TB with no monthly minimum
• Account-state history: the Solana Account Archive answers getAccountInfo for any account at any historical slot back to July 2025, with backfill extending coverage further back over time. Backfilled ranges are re-executed through the validator runtime of the corresponding era and verified against the network's own consensus artifacts.
• Enhanced account data: up to 1,000 token accounts (50x more) in a single getTokenLargestAccounts call
• Multi-region, enterprise-grade infrastructure with autonomous failover and fault-tolerant gRPC streaming that survives node failures without dropping data

Alongside RPC and streaming, we ship gas sponsorship, enriched data APIs, webhooks, request logs, and real-time analytics. gRPC migration is a URL change, Yellowstone-compatible across Rust, TypeScript, Go, and anything that compiles a .proto. Solana teams can also apply to the $20M Solana Fund for up to $25k in credits.

• Strengths: 99.99% uptime, lowest average latency in the global, US East, and US West benchmark views, genesis-depth history plus account-state archive, up to 10-20x faster heavy and archive calls, usage-priced Yellowstone-compatible gRPC, 30M CU/month free tier, transparent pricing, full developer suite, multi-chain.
• Best fit for: Teams running mission-critical Solana workloads (wallets, exchanges, payments, indexers), teams that read from Solana and other chains through one provider, and teams that want a full developer suite (gas sponsorship, webhooks, analytics, enriched APIs) under one roof. Teams serving mostly EU or Asia-Pacific users can confirm regional latency on the free tier before committing.
• Pricing: Free 30M CU/mo · PAYG $0.52/1M CU · gRPC from $75/TB (no monthly minimum) · Enterprise custom.

2. Helius: Solana-native specialization

Helius focuses exclusively on Solana. That focus pays off in enhanced APIs for NFTs, tokens, and transaction parsing (including a getTransactionsForAddress method for full parsed history per address), genesis-depth transaction history on every plan, and hosted validator and dedicated-node options. LaserStream, its gRPC streaming service, runs on devnet on all plans; mainnet access starts on the Business plan at $499/month, with replay of up to 48 hours.

For teams evaluating alternatives to Helius, Alchemy is the closest match on the core Solana surface (RPC, genesis-depth history, Yellowstone-compatible gRPC, webhooks) with three concrete differences: gRPC is usage-priced at $75/TB with no plan minimum instead of gated behind a $499/month tier, the Account Archive adds account-state history that transaction-history archives don't cover, and one account covers Solana plus 100+ other chains. In our October 15, 2026 benchmark, Helius averaged 38.29 ms globally against our 8.07 ms. The benchmark measures plain read methods only, so test both against your real traffic.

• Strengths: Solana depth, transaction landing, enhanced parsing APIs, direct leader connections, genesis-depth history.
• Focus: Solana-only teams, enhanced NFT and token APIs, and validator services.
• Pricing: Free 1M credits / 10 RPS · Developer $49/mo · Business $499/mo (LaserStream mainnet starts here) · dedicated nodes from $2,900/mo.

3. QuickNode: broad multi-chain footprint

QuickNode is one of the oldest RPC vendors, now spanning 80+ chains across 130+ networks with 17+ regional deployments, built-in analytics, and a credit-backed 99.99% uptime SLA. It documents full Solana historical data, and Yellowstone gRPC is available as a $499/month add-on on lower tiers (included from the Scale plan up).

• Strengths: Multi-chain coverage, analytics, credit-backed SLA, marketplace add-ons.
• Focus: Multi-chain teams that value breadth and built-in analytics.
• Pricing: Free 10M credits · Build $49/mo · Scale $499/mo (includes gRPC).

4. Triton: bare-metal, usage-priced infrastructure

Triton is purpose-built for HFT, MEV, and trading infrastructure, running bare metal across NA/EU/APAC with GeoDNS routing. It moved to pay-as-you-go pricing with a $125 minimum deposit, then $0.08/GB bandwidth plus $10 per million calls for RPC, $0.08/GB for streaming, and shred streaming from $900/month per IP per data center. Its open-source Old Faithful project archives Solana history from genesis, and historical calls route to it. Triton now covers Sui and Monad alongside Solana.

In our October 15, 2026 benchmark Triton averaged 17.89 ms in US East. Its global average of 103.82 ms is pulled up by EU Central and AP Southeast, and the benchmark covers read methods only, not the shred streaming and landing workloads Triton is built for.

• Strengths: Latency in its home regions, bare-metal infra, usage pricing with no monthly minimum, genesis archive via Old Faithful.
• Focus: HFT, MEV, and other latency-sensitive workloads that want dedicated-grade infrastructure billed by usage.
• Pricing: PAYG with $125 minimum deposit · $0.08/GB + $10/M calls for RPC · $0.08/GB for streaming · shred streaming from $900/mo per IP per data center.

5. Ankr: decentralized infrastructure network

Ankr serves 75+ chains with global anycast routing and a large free tier (200M API credits/month). The tradeoff on Solana is history. Ankr's own Solana docs state its nodes retain roughly the last 16 hours of transaction history, and recommend a separate archive provider for anything older.

• Strengths: Cheap entry, multi-chain, distributed architecture, large free tier.
• Focus: Cost-sensitive multi-chain teams reading recent state, without deep-history or Solana-specific tooling needs.
• Pricing: Free 200M credits/mo · PAYG $0.10 per 1M credits · subscription deals from $500/mo.

6. Syndica: Solana-specific, compliance-oriented

Syndica focuses solely on Solana with RPC nodes, its ChainStream API, and analytics, and it builds Sig, an open-source Solana validator client. Pricing has moved to a sales-led model. A 10M requests/month free tier is public, and paid plans are quoted through its enterprise calculator rather than a published price list.

• Strengths: Solana-only focus, observability, validator-client engineering depth.
• Focus: Solana-only enterprise teams comfortable with a sales-led motion for RPC, streaming, and observability.
• Pricing: Free 10M req/mo · paid plans custom (sales-led).

7. Chainstack: flexible infrastructure tiers

Chainstack offers three deployment models (geo-balanced global nodes, managed dedicated nodes, and self-hosted nodes) plus full-history Solana archive access starting at $49/month on the Growth plan. Yellowstone gRPC with Jito ShredStream is a Marketplace add-on available from the Growth plan: $49/month for one stream, $149/month for 5, and $449/month for 25, on top of the plan fee, so the entry point is about $98/month.

• Strengths: Deployment optionality, low-cost archive and gRPC entry, compliance support.
• Focus: Teams that want to choose between shared, dedicated, or self-hosted deployment models and are comfortable making the configuration choices that come with them.
• Pricing: Free 3M RU/mo · Growth $49/mo (20M RU) · Pro $199/mo · custom SLAs on Enterprise.

8. dRPC: multi-provider aggregator

dRPC routes across many underlying providers from a single endpoint, covering 130+ chains across 224 networks with live per-node status. It charges the same for full and archive nodes, and Yellowstone gRPC for Solana is available as a marketplace add-on. Useful as a redundancy layer rather than a primary.

• Strengths: Aggregated routing, transparency dashboard, uniform archive pricing.
• Focus: Teams using dRPC as a routing or redundancy layer across multiple upstream providers, with active monitoring of the underlying pool.
• Pricing: Free 210M CU/mo (public nodes) · PAYG $6 per 1M requests · SLA on Enterprise.

9. Blockdaemon: institutional infrastructure

Blockdaemon serves 400+ institutions and secures $110B+ in assets. On Solana it offers archival nodes with complete historical data, event streaming, and staking APIs, with RPC sitting inside a broader custody and staking product suite. Plan pricing is custom.

• Strengths: Institutional trust, MPC wallets, staking integration.
• Focus: Institutional and enterprise teams whose RPC needs sit alongside staking, custody, or MPC requirements.
• Pricing: Free 3M CU · plans custom (published overage rates from $0.00002/CU).

Solana gRPC streaming pricing compared

If you need a Yellowstone-compatible stream, the biggest difference between providers is how it's sold. Alchemy's $75/TB is the lowest per-TB or per-GB rate among the providers below that publish one (Triton lists $0.08/GB, about $80/TB), with no monthly minimum. The other providers sell streaming as flat fees, which can cost more at low volume and less at high volume. Prices below come from each provider's public pricing page as of October 15, 2026.

Streaming pricing changes often, and bandwidth-priced and flat-fee plans cross over at different volumes, so estimate your monthly data volume before choosing. Our Solana gRPC launch post walks through the setup.

Which Solana gRPC providers support server-side filters, replay, and production SLAs?

Beyond pricing, there are a few additional capabilities often used by developers.

Server-side filters

Server-side filters help cut costs and the amount of data you handle, because you only receive and pay for what your app needs.

The basic Yellowstone filters cover three jobs. On transactions, you can include, exclude or require specific accounts, and skip vote or failed transactions. On accounts, you can match by address, by owning program, or by bytes in the account data.

Yellowstone also has newer filters. Cuckoo filters (Triton) track very large account lists. The token-account option (Helius calls it tokenAccounts) also catches token transfers into a wallet. Data slicing sends only the bytes you ask for from each account update. Which of these you can use depends on what your provider has turned on.

Filter limits are configurable, so they vary by provider. Check how many account filters a subscription can have and how large a memcmp filter can be.

Replay after a disconnect

Providers differ on this more than on any other capability here, and most teams only find out they need it after their first outage.

Replay means the provider resends what you missed once you reconnect. Providers handle it in one of two ways:

• Short buffer: 100 to 150 slots, about 40 to 60 seconds. It covers a brief network blip and nothing longer.
• Long window: hours of history. You need this to get through a deploy, a pod restart or a cloud networking problem, because those all last longer than a short buffer.

If your gap is longer than the provider's window, the data you missed is gone unless you've built your own backfill. Pricing pages often make the two sound alike, so check which one you're buying.

On Alchemy:

• Set from_slot on the SubscribeRequest to the last slot you saved successfully.
• The stream sends everything from that slot until it catches up to the present.
• The window is 432,000 slots, about 48 hours.
• Work out the slot when your app starts instead of hard-coding one. A fixed slot eventually falls outside the window as the chain moves on.
• For anything older than 48 hours, you need a historical backfill.

Production SLAs for streaming

Providers often don't extend an uptime SLA to streaming on regular paid plans. There are two ways to judge reliability before you commit:

• Check historical uptime. Our Solana benchmarks show how different providers have performed over time.
• Ask about an enterprise plan. Enterprise plans can include an SLA for production workloads at scale. We help enterprises like Circle and Phantom run on Alchemy. Talk to our team for enterprise plans to talk through contractual terms.

With any provider, get the SLA terms in writing and ask whether they cover streaming or only standard request-response traffic.

What is the best Solana gRPC streaming provider for real-time data?

There's no single best provider. The right choice depends on which problems you most need to avoid. These are the factors to compare:

• Affordability. Look at how pricing is structured, not just the headline number. Flat plans and per-stream add-ons cost the same whether you stream constantly or in bursts, and many cap how many subscriptions you can run at once. If your volume is flat and predictable, a fixed monthly plan may cost less than usage-based pricing.
• Replay. If your stream drops, can you get back what you missed? Check whether a provider offers a true backfill or only a short reconnect buffer, which usually covers about a minute.
• Filters. Check which filters your provider supports beyond the basic Yellowstone set, such as Cuckoo filters or token account matching. The more precisely you filter, the less data you pay for.
• Slow nodes. Ask how each provider handles a node that stalls instead of failing outright. A stalled node can hold up your stream without ever triggering an error.

Alchemy covers all four:

• Affordability: We charge $75 per TB, with no plan prerequisite, no monthly minimum and no limit on streams. A bursty MEV bot and a steady indexer pay the same rate.
• Replay: We replay up to 432,000 slots, about 48 hours, with no separate pipeline to run.
• Filters: We support the standard Yellowstone filters, including data slicing, so you only receive the bytes you need.
• Slow nodes: We spread every subscription across multiple upstream Solana nodes and use the fastest healthy response, so a GC pause, a node falling behind or a restart doesn't show up in your stream.

When should you switch RPC providers?

Switch when any of these is true for more than a sprint:

• You're hitting rate limits or 5xx errors during normal peak traffic.
• Your provider can't return the history you need for compliance, analytics, or debugging, whether that's old transactions or account state at a past slot.
• Latency P95 from your user regions has drifted above your product SLA.
• Your bill is unpredictable month-to-month and you can't model the next quarter.
• A feature you depend on (gRPC streaming, archival access) is gated behind a plan tier priced far above your usage.

Migration itself is simple. Solana uses standard JSON-RPC, so most switches are an endpoint URL and API key swap. Most teams stage it:

• Pre-stage the new provider in staging and replay a week of production traffic.
• Drop DNS TTL under 30 seconds.
• Roll traffic gradually (1% → 10% → 50% → 100%) with health checks at each step.
• Keep the old provider warm for 48 hours in case of rollback.

Frequently asked questions

What is the best Solana RPC provider in 2026?

For mission-critical production applications, Alchemy is the best Solana RPC provider in 2026. It offers 99.99% uptime, multi-region failover, the lowest average latency globally, in US East, and in US West on its public benchmarks (October 15, 2026 snapshot), up to 10x faster heavy reads and 20x faster archive calls, account-state history through the Solana Account Archive, and Solana plus 100+ other chains under one account. Its Yellowstone-compatible gRPC starts at $75/TB with no monthly minimum, the lowest usage-based streaming rate among the providers compared here, and there is a 30M CU/month free tier. It already powers Phantom, Solflare, Robinhood, OpenSea, and Circle.

Which Solana RPC provider is the fastest?

It depends on the region. In Alchemy's public benchmark for the 24 hours ending October 15, 2026, Alchemy had the lowest average latency globally (8.07 ms), in US East (7.40 ms), and in US West (10.76 ms). In EU Central and AP Southeast, averages across the leading providers were within a few milliseconds of each other, and the regional table above lists every result. The benchmark covers seven read methods from four AWS regions and refreshes every few minutes, so check the live dashboard and test your own traffic before deciding.

What is the best Solana RPC alternative to Helius?

Alchemy is the most common Solana RPC alternative to Helius. It covers the same core surface (RPC, genesis-depth history, Yellowstone-compatible gRPC, webhooks) with usage-priced gRPC at $75/TB instead of a $499/month plan floor, account-state history back to July 2025 through the Account Archive, and 100+ chains under one account. Live latency comparisons between the two are published on Alchemy's public benchmarks page.

Which Solana RPC provider has the fastest archival data?

Alchemy serves archive calls up to 20x faster per its published Solana infrastructure benchmarks, and its Account Archive answers getAccountInfo at any historical slot back to July 2025, which transaction-history archives don't cover. For genesis-depth transaction and block history alone, Helius, QuickNode, Triton (via Old Faithful), and Chainstack also document full coverage.

How much does Solana gRPC streaming cost?

It varies by provider. Alchemy charges $75/TB pay as you go with no monthly minimum, the lowest per-TB or per-GB rate among the providers compared here. Helius LaserStream mainnet starts on the $499/month Business plan. QuickNode sells Yellowstone gRPC as a $499/month add-on on lower tiers and includes it from the Scale plan. Triton bills streaming at $0.08/GB, and Chainstack sells Yellowstone gRPC as an add-on from $49/month for one stream, on top of its $49/month Growth plan. Compare against your expected monthly data volume, because per-GB and flat-fee pricing cross over at different points.

Which Solana gRPC providers support server-side filters, replay, and production SLAs?

All the providers compared here support the standard Yellowstone filters. Triton and Helius document extras on top of that. Triton supports Cuckoo filters for tracking very large account lists, and Helius has a tokenAccounts filter that also catches token transfers into a wallet. Alchemy supports the standard filters, including data slicing, so you only receive the bytes you need.

Replay varies the most. Some providers offer a short reconnect buffer that covers about a minute, and others keep hours of history. Alchemy replays up to 432,000 slots, about 48 hours, using from_slot.

Providers often don't extend an uptime SLA to streaming on regular paid plans. Enterprise plans can include one, so contact sales if you need contractual uptime terms. To see how our infrastructure has performed, check our Solana benchmarks.

What is the best Solana gRPC streaming provider for real-time data?

There's no single best provider. Compare on affordability, replay, filters and slow nodes. On affordability, flat plans and per-stream add-ons cost the same whether you stream constantly or in bursts, and many cap how many subscriptions you can run. On replay, check whether a provider offers a true backfill or only a short reconnect buffer. On filters, check which ones go beyond the basic Yellowstone set, because tighter filters mean less data to pay for. On slow nodes, ask how a provider handles a node that stalls instead of failing outright.

Alchemy covers all four. We charge $75 per TB with no plan prerequisite, monthly minimum or limit on streams. We replay about 48 hours. We support the standard Yellowstone filters, including data slicing. We also spread every subscription across multiple upstream Solana nodes, so a stalled node doesn't show up in your stream.

What's the difference between public and private Solana RPC endpoints?

Public endpoints are free but capped at 100-200 requests/second per IP, have 2-5 second data delays, no SLA, and shared resources that suffer "noisy neighbor" effects. Private endpoints offer 1,000+ requests/second per API key, real-time data, 99.9%+ uptime SLAs with financial penalties, and dedicated resources. Use public for prototypes; use private for anything serving real users.

How do I choose a Solana RPC provider?

Work down a 7-point checklist: (1) uptime SLA ≥99.9%, (2) P90 latency tested from your user regions under 10x peak load, (3) archival depth matching your compliance and analytics needs, (4) transparent CU-based pricing, (5) websocket reliability over 95%, (6) bundled tooling (webhooks, logs, alerts), and (7) migration friction. Disqualify on the first failure rather than scoring everything.

When should I switch Solana RPC providers?

Switch when you're hitting rate limits at normal peak, your provider lacks archival history you need, P95 latency drifts above your SLA, billing is unpredictable, or a feature you depend on is gated behind a plan tier priced far above your usage. Solana's standard JSON-RPC interface makes most migrations an endpoint and API key swap. Plan a staged rollout with DNS TTL under 30 seconds and a 48-hour rollback window.

What free tiers do top Solana RPC providers offer?

Free tiers use different units, so compare against your real call mix. Alchemy offers 30M CU/month with the full feature surface, including archival access and webhooks. QuickNode offers 10M credits, Helius 1M credits at 10 RPS, Chainstack 3M request units, Syndica 10M requests, and Blockdaemon 3M CU. Ankr's 200M credits and dRPC's 210M CU tiers are larger on paper but run against shared or public node pools.

What are compute units (CUs) and how do they affect Solana RPC pricing?

Compute units represent the computational cost of an RPC request. Simple queries like getBalance cost ~20 CU; heavier ones like getProgramAccounts cost more. CU-based pricing (for example, Alchemy's $0.52 per 1M CU) gives more predictable bills than per-request pricing, because traffic mix doesn't distort cost. Always normalize competitor pricing into a per-million-CU equivalent before comparing.

Can I migrate Solana RPC providers without code changes?

Usually yes. Solana follows a standard JSON-RPC interface, so most migrations require only an endpoint URL and API key swap. Verify any provider-specific extensions (enhanced APIs, gRPC streams, parsing helpers) are available on the new provider, and test in staging before redirecting production traffic.

When does self-hosting Solana RPC nodes make sense?

Almost never on cost alone. Self-hosting means $5,000+ hardware per node, 1TB+/mo bandwidth, 24/7 ops coverage, and constant tuning against Solana's rapidly evolving runtime. SaaS providers exist because the infrastructure is genuinely hard. Self-host only when regulatory or sovereignty requirements force the issue.

Ready to build on Solana? Get started with Alchemy's Solana infrastructure with 30M free CUs a month, no contract and no minimum commitment, and apply to the $20M Solana Fund for up to $25k in credits.