Contention-aware fees
Fee data comes from hot accounts such as AMM pools and Jito tip accounts, so the number reflects what competitive transactions truly pay.
6ua4jX4BuwoEMAfEhXq9WLyZHrfraatLMBjwfzyYpumpExecution intelligence for Solana
Agents, perp traders and tokenized-asset apps lose money to dropped transactions. Axiomora reads the live fee market and Jito tip floor, prices your compute budget to the percentile you need, and hands it to your bot over a single API call.
Live telemetry
Current slot
Non-vote TPS
SOL price
Jupiter price API
Epoch
Fee needed to write to contended accounts (top AMM pools, Jito tip account) over the last 150 slots, oldest on the left.
How it works
Four deterministic steps, every one of them verifiable against public Solana RPC data.
Axiomora samples recent prioritization fees on the most contended writable accounts, the ones real traffic fights over, instead of the near-zero global average that tells you nothing about competitive inclusion.
getRecentPrioritizationFees / contended accounts / 150 slots
Fees are reduced to percentiles and combined with the Jito landed-tip floor. Each urgency tier maps to a percentile, so you choose how much certainty to buy instead of guessing a number.
p50 / p75 / p90 / p99 + Jito tip floor
The exact transaction you intend to send is simulated on mainnet. The compute limit is sized from real units consumed with a safety margin, so you never overpay for unused compute.
simulateTransaction / units consumed / margin
A compute-budget transaction is built in your browser and signed by your own wallet. Axiomora never sees your keys, and the signature is tracked until the cluster confirms it.
ComputeBudgetProgram / wallet signature / confirmation
Fee planner
Pick a workload and an urgency. Axiomora prices the compute budget from the last 150 slots, simulates it against mainnet to size the limit exactly, then signs and lands it with your wallet.
Top decile. For swaps and order placement.
A SOL transfer wrapped in your priced compute budget. Leave the recipient empty for a fee-only self-transfer.
Urgency tiers
Live values for each tier are shown in the planner above. Use this guide to choose the right one.
| Tier | Use it for | Trade-off |
|---|---|---|
| Economy | Non-urgent settlement, batch jobs, anything that can retry | Lowest cost, may wait several slots |
| Standard | Everyday transfers, swaps in calm markets, treasury moves | Balanced cost and inclusion |
| Fast | Volatile markets, perp orders, time-sensitive agent actions | Higher percentile plus Jito tip |
| Critical | Liquidations, arbitrage, large tickets where a miss is costly | Top percentile and tip, maximum certainty |
Capabilities
Fee data comes from hot accounts such as AMM pools and Jito tip accounts, so the number reflects what competitive transactions truly pay.
Landed-tip percentiles from the Jito block engine feed the Fast and Critical tiers, pairing compute price with a sized bundle tip.
Compute limits come from a real mainnet simulation of your transaction, not a flat 200k default that wastes fee budget.
Transactions are built client-side and signed by Phantom, Solflare or Backpack. No keys, no deposits, no accounts.
One CORS-enabled GET returns the exact compute-budget values a bot attaches to its transaction. No SDK or API key required.
RPC traffic runs through an allow-listed server proxy with input validation, so credentials and unsafe methods never reach the browser.
Agent-ready API
The same engine behind the planner is exposed as an open JSON endpoint with CORS enabled. Autonomous agents, keepers and trading bots call it right before signing and receive the exact compute-budget instructions to attach.
const fees = await fetch("https://YOUR_DOMAIN/api/fees?urgency=fast&cu=200000").then(r => r.json())
tx.add(ComputeBudgetProgram.setComputeUnitLimit({ units: fees.computeUnits }))
tx.add(ComputeBudgetProgram.setComputeUnitPrice({ microLamports: fees.computeUnitPrice }))GET /api/fees?urgency=fast&cu=200000// Response appears here
Use cases
Every serious Solana workload shares one risk: a transaction that does not land when it should.
Agents that pay, trade and rebalance without a human cannot retry blindly or overpay.
Call the API before every signature, cap spend per action in code, and fall back to a lower tier when the market is calm.
Onchain orderbooks reward the transaction that lands first, and a dropped order is a missed fill.
Run Standard in quiet markets and escalate to Fast or Critical only when fees spike, so you pay for speed only when it matters.
Settlement for tokenized equities and stablecoins cannot tolerate silent drops on large tickets.
Use Standard by default and price certainty by ticket size, with a simulation check before anything is signed.
Liquidation races are decided inside a single slot.
Critical tier pairs a top-percentile compute price with a Jito tip sized from landed bundles.
Mint windows create sudden contention that pushes normal fees below the inclusion line.
Read live contended-account fees right before the window opens and set your budget from real data.
Users blame the app when a transaction stalls, and a fixed fee is either wasteful or too low.
Show users a clear Economy, Standard or Fast choice backed by live data instead of one hardcoded fee.
FAQ
No. Transactions are built in your browser and signed by your own wallet. Axiomora has no accounts, deposits or custody of any kind.
No. Tiers are statistical estimates from recent network data. They raise the odds of inclusion but cannot guarantee it, because conditions can change within a single slot.
Solana RPC for slots, performance samples and prioritization fees, the Jito block engine for landed-tip percentiles, and Jupiter for the SOL price.
No. The fee endpoint is a plain CORS-enabled GET request. For heavy production traffic, deployers should point the app at a dedicated RPC provider.
The default action is a self-transfer that costs only the network fee plus your chosen priority fee, a few thousand lamports. You can simulate first at no cost.
Phantom, Solflare and Backpack, using their injected browser providers.
Price your next transaction against live mainnet data, or wire the API into your agent in a single call.