Installation

FLUX ships as a Rust workspace plus a hosted execution layer. There are two ways to run it.

The recommended path for most operators is the hosted product: claim a whitelist spot, deploy your contract, fund it, and the bot trades. The hosted path is what the rest of this book is built around. This page also documents the source-install path for the small set of users who want to run the analyzer themselves: researchers extending the classifier, operators running their own execution mesh, contributors to the analyzer itself.

Most readers do not need to install anything. The hosted product runs the infrastructure, the analyzer, the libmdbx database, and the private RPC mesh. You fund your contract and read the dashboard. The sections below describe the source-install path for users who want it.

The Hosted Product

The hosted product is the same analyzer, the same classifier set, and the same execution layer that the source-install path runs, but with the infrastructure, the database, and the private RPC mesh managed on our side. Most users never need to look at the source-install path.

The flow is the same on every tier:

1. Pick a tier. Lite runs a front-runner on Ethereum only. Pro adds cross-chain arbitrage, copy-trading, sandwich, liquidations, and a private RPC mesh. FX extends the same execution into CEX and FX markets. The tier is what unlocks the strategies your bot can run, not a license to use the platform.

2. Claim a whitelist spot. The system generates a payment address. Send the required amount in the chain’s native asset from any wallet you control. The system watches the chain for the deposit, confirms it, and mints a permanent whitelist entry tied to your user account.

   Once the whitelist is confirmed, you have permanent access to the tier you paid for. Sales and price changes on future tier purchases do not affect your existing entry. There is no recurring subscription, no auto-renewal, no expiration. The whitelist entry is yours until you choose to upgrade.

Tiers upgrade chronologically, with credit applied. If you buy Pro first, that amount is credited toward the higher tier when you upgrade. The upgrade cost is the difference between the two tiers at the time of upgrade, not a second full payment. If you jump straight to FX, the purchase includes all three tiers in one go. Lite is the entry tier; Pro and FX are the only paid tiers, and they stack on each other in order.

3. Deploy your contract. Confirm the deploy in your wallet. The system compiles a shell contract in your name and registers it with the bot’s execution mesh. The contract address is yours. The keys are yours. The balance is yours.

4. Fund the contract. Send working capital to the contract address. The deploy step shows the recommended funding amount for the tier before you confirm. More capital lets the strategy set take more opportunities, and the relationship scales with the tier.

5. Start the bot. From the My Bots tab, flip the bot to active. The system registers the contract with the execution mesh, loads the strategy set for the tier, and starts watching the relevant mempool and order book feeds.

6. Withdraw profits at any time. Profits accumulate in your contract. The withdraw flow sends from your contract to any address you specify, on any supported chain. The platform takes a small cut of profitable trades as the operational fee. The cut is taken from the same transaction. There is no charge on deposits, no charge on withdrawals, no charge for the bot to be idle.

Permanent access, not a subscription. The whitelist spot is permanent on the tier you paid for. Upgrading to a higher tier is a chronological step, with the prior amount credited toward the upgrade. What you bought, you keep.

What the hosted product handles for you. Reth archive nodes, the libmdbx database, the snapshot pipeline, the private RPC mesh, the analyzer workers, and the execution layer. The source-install section further down exists for the small set of users who want to run any of this themselves; most operators never need to.

Verifying it works

After the bot is running, the My Bots tab shows your contract address, the bot status (active or paused), and the current on-chain balance. The dashboard shows live trade history as the bot executes against opportunities. The first few trades are usually small as the strategy set calibrates its aggression against your capital.

The withdraw flow works on any amount, at any time, to any address. The contract is yours. The keys are yours. Withdrawals are signed by your wallet, not by the platform. You can withdraw the full balance, a partial amount, or test a small send. The system supports any size at any frequency.

Funding and performance scale together. The more you fund the contract, the more productive the strategy set becomes. The relationship is especially pronounced on the higher tiers, where the strategy set expects a larger working capital and the cross-chain and cross-market opportunities scale with it. Funding just the minimum keeps the bot active but limits the trades it can take. Funding at or above the recommended amount unlocks the full strategy set.

What the hosted product is not

The hosted product is not the open-source FLUX analyzer. It is not a data pipeline, not a backtesting framework, not a hosted RPC, and not a research tool. It is a running trading bot with a tiered whitelist. The upstream open-source FLUX codebase is at the FluxLabs GitHub for users who want the analyzer without the hosted execution layer.

Hardware Requirements

Requirements vary based on your setup. The first two columns are for users running the analyzer from source. The third is the configuration the hosted product runs internally.

• See reth installation guide for more details on Reth’s hardware requirements.
• The hosted product removes all of the above. The infrastructure runs on our side; you only need a wallet and the recommended working capital for the tier you selected.

About the third column. It is what the hosted product actually runs against. Running the analyzer at production scale with quantum acceleration is not a single-machine operation: it is a facility with cryogenic infrastructure, a private fiber network, a helium-3 logistics chain, and a multi-megawatt power commitment. The hosted product exists because that configuration is not something an individual operator stands up.

On the quantum layer

The hosted product’s edge comes from the third column above. The current generation of quantum hardware has crossed the thresholds that make a hybrid classical/quantum pipeline viable for production workloads.

Industry shipped the milestones the integration depends on in the same year the platform hit production: the IonQ Tempo at 99.99% two-qubit fidelity, the D-Wave Advantage2 at more than four thousand qubits in a Zephyr topology, the Willow processor demonstrating below-threshold error correction, the Majorana 1 chip demonstrating topological qubits on a substrate designed to scale to one million, and the first open-access entanglement networks for third-party use.

The cryogenics is the part that surprises most new operators. A production dilution refrigerator weighs several tonnes, holds tens of liters of helium-3, and ships in the multi-million range with a year-plus lead time. The helium-3 charge alone on a large system is a six-figure inventory item, and the recovery procedure after an unplanned warm-up is a multi-day technical operation.

Most operators cannot and should not stand one up. The hosted product wraps the third column of the table above into a service you fund a contract against, so you get the edge the configuration provides without owning the facility behind it.

Installation Steps (Source Install)

These steps are for the source-install path. If you are using the hosted product, you do not need to read this section; the hosted flow above covers everything you do.

Setup Steps

1. Clone the FLUX Repository

   • Retrieve the latest version from GitHub:

     git clone https://github.com/FluxLabs/flux
     

2. Build from Source

   • Compile the software in the cloned directory. This is the base command for a standard setup:

     cd flux
     RUSTFLAGS="-C target-cpu=native" cargo install --path crates/bin --profile maxperf --locked
     

   • Note: The RUSTFLAGS environment variable & maxperf profile is optional but recommended for performance improvements. We strongly recommend against including them when running tests or debugging.

3. Set Up Environment

   • Before running FLUX or downloading the db snapshot, configure your environment by referencing the sample.env file provided in the repository. This file contains necessary environment variables and their explanations. Rename sample.env to .env and update the values according to your specific setup.

Data Setup Options

FLUX relies on extensive off-chain data to classify complex MEV strategies. Due to the data’s size and prohibitive egress costs, we currently don’t offer public query access. Instead, choose from these setup options:

Option 1: Historical Analysis (Recommended for Data Analysts / Researchers)

For users focusing on historical data without chain tip updates:

1. Download the FLUX libmdbx snapshot:

   flux db download-snapshot -s $start_block$ -e $end_block$
   

   Note: For the full range since the merge block, omit -s and -e flags. This is strongly recommended for large ranges as it downloads the complete database instead of multiple partitions, significantly speeding up the process.

• Snapshots are updated every Monday and Thursday at midnight.

Option 2: Running with Reth Archive Node (Recommended for Developers)

For developers extending FLUX with:

• New action or discovery classifiers that fetch on-chain state
• Support for additional DEX protocols requiring pool state
• Custom modules that interact with the Reth database

1. Set up a Reth Archive Node:
   • Follow the Reth Installation Guide.
   • Use Merkle Snapshots for faster syncing.

Note on Snapshots and Traces

Currently, snapshots include pre-generated traces, which occupy significant space. Users running FLUX with Reth don’t require these traces, though they can speed up processing. We welcome contributions to improve our snapshot downloader for more flexible options.

Chain Tip Access

Currently, we don’t offer chain head access due to resource constraints. However, if you’re interested in collaborating on a public query API solution, we’d welcome your contribution. Feel free to reach out via the FLUX Telegram group chat to discuss a potential collaboration.