Corvus Labs

Pricing

Corvus plans compared by request capacity (RPS), transaction capacity (TPS) and which interfaces each one unlocks.

A plan is capacity: a request allowance (RPS), a transaction submission allowance (TPS), and a set of interfaces. You rent in the Corvus dashboard, from the presets below or by setting your own RPS and TPS on the Custom plan.

Starter

For smaller production workloads

$160/month

$0.22/hour effective

RPC RPS
200
Submit TPS
50
  • RPC
  • gRPC
  • aRPC

Standard

Balanced throughput and cost

$325/month

$0.45/hour effective

RPC RPS
500
Submit TPS
125
  • RPC
  • gRPC
  • aRPC

Custom

Dial in any capacity

Quoted

Priced on the capacity you choose

RPC RPS
Your choice
Submit TPS
Your choice
  • RPC
  • gRPC
  • aRPC

Choose any RPS and TPS, including capacity above the presets. Configure it and see your price in the dashboard.

Monthly billing controls. Effective hourly rates use a 720-hour month and are rounded to two decimals. Custom capacity is configured and priced in the dashboard.

RPS and TPS are separate allowances that don't convert into each other. sendTransaction costs 0 RPS units and draws on TPS instead.

What a plan includes

PlanRPSTPSJSON-RPC & WebSocketYellowstone gRPCaRPC
Starter20050
Standard500125
Professional1,000250
CustomYour choiceYour choiceWhen RPS + TPS ≥ 500

aRPC unlocks when RPS and TPS together reach 500. The same total applies on a Custom plan, so size with that in mind if you want streams.

Falcon sits outside this table: its TPS allowance attaches to your API key rather than to a plan, and every transaction pays a tip. See Falcon.

An NFT pass is the other way in. Instead of renting one region, a pass carries a fixed high allowance across every region for a monthly renewal. You never need one to rent.

Whichever you pick, you also get:

  • One region, chosen when you rent, from the five Solana RPC metros: Frankfurt, Amsterdam, London, New York and Tokyo. See Regions.
  • IP allowlist access on the standard regional endpoints; register a public egress IP in the dashboard. See Access & authentication.
  • Support on Discord for API keys, tokens and anything the dashboard doesn't cover.

Call an interface your plan doesn't include and you get HTTP 403, code -32003, prefix You have no access to this service.

What RPS buys

RPS is metered by weighted method cost rather than request count, and every item in a JSON-RPC batch is counted on its own. On a 500 RPS plan:

WorkloadCost per callCalls per second at 500 RPS
Ordinary reads (getSlot, getBalance, …)1500
getAccountInfo2250
getMultipleAccounts, 100 accounts2002.5
getTransaction1050
getProgramAccounts30~16

getMultipleAccounts is the row that moves: below 800 plan RPS it costs 2 units per account, and from 800 RPS it drops to ceil(accounts / 20), so the same 100-account call on Professional costs 5 units. NFT pass holders get that cheaper rate on any plan. The formula has all three cases.

Six scan-style methods also carry a concurrency limit, independent of RPS, so a getProgramAccounts-heavy workload can hit Scan request timed out waiting for concurrency slot with RPS headroom to spare. The full rules are in Solana RPC rate limits. These weights apply to the shared regional JSON-RPC endpoints.

If your mix doesn't fit any of the presets, see Custom deployments.

How billing works

Plans bill monthly in USD. The hourly figure on each card is for comparison, not a separate metered price:

effective hourly rate = monthly price / 720 hours

It's rounded to two decimals.

Custom isn't a quote request. You set the RPS and TPS you want, including capacity above the Professional preset, and the dashboard prices that configuration before you commit.

Compare providers

Corvus sells capacity, Helius combines plans with credits, and Triton publishes usage-based pricing. The normalized RPC cost comparison puts all three on one axis, with the assumptions, exact values and source links.

Pricing verified July 25, 2026.

On this page