The generated installer command writes the visitor's own BTX payout address into the miner config.
Customer mining status
Your mining path is the main workflow.
Treasury details stay public, but below the mining flow.
Each payout address starts with a 7-day fee-free trial. After that, the backend target fee is 0.50% through public payout accounting.
No account, email, custody, or bot is required to generate the miner command. The page does not store visitor input.
New installs target stratum.drinknile.com:3333 so there is no later migration from the old pool.
Watch accepted shares, GPU utilization, expected BTX/hour, worker identity, and aggregate network stats.
A full BTX Start per-wallet web dashboard still needs our own share, payout, and balance index.
Quick start
One address. One command. No account.
Paste your own BTX payout address, copy the generated command, and run the preflight on a Linux machine with an NVIDIA GPU. The installer is already wired to the first-party BTX Start stratum target and will fail loudly until that backend is reachable.
Your payout address is inserted into the command locally in your browser.
Run one command. Worker name is optional and defaults to a stable label.
The miner log and GPU view show whether the rig is actually working.
Pool payouts are batched weekly for non-zero balances.
No account, email, chat app, wallet connection, or sales call.
The command mines to the BTX address you paste. BTX Start does not hold user funds.
The installer uses the prebuilt solver path and verifies the binary hash.
First 7 days are 0.00%. After that, the disclosed target fee is 0.50% and must route to the public treasury wallet.
Install for BTX Start pool
Downloads the miner package, verifies the solver hash, writes config for stratum.drinknile.com:3333, and runs the CUDA engagement smoke test.
curl -fsSL https://drinknile.com/install.sh | bash -s -- --address 'btx1z...YOUR_BTX_ADDRESS...' --worker 'default'
Optional dry run
Runs host, GPU, artifact, address, and network checks without writing config or installing packages.
curl -fsSL https://drinknile.com/install.sh | bash -s -- --preflight --worker 'default'
Laptop first
Start with the machine you already own.
No rented GPU is required to learn the flow.
Run preflight first. If the laptop has working NVIDIA drivers and reaches the BTX Start stratum endpoint, use the same one-line install command as any other rig.
Use Ubuntu under WSL2 with current NVIDIA drivers. The miner path is still Linux-based, so verify nvidia-smi inside WSL before installing.
The public one-line miner path is not a profitable CPU-only or Apple Silicon path yet. Use the laptop to read the repo, run preflight, prepare the wallet, and understand the logs before scaling.
Prepare your payout address
Use the same BTX address field above. The generated local commands mine only to the address you paste; BTX Start does not custody funds.
Run a readiness check
Preflight tells you whether your laptop can actually mine: operating system, Python, release artifact, NVIDIA visibility, payout address format, and stratum reachability.
Install only if preflight passes
A laptop that cannot expose CUDA should not be pushed into paid mining. Treat it as a learning machine until a supported GPU path is available.
Watch the miner locally
Accepted shares in the log and sustained GPU load are the two signals that matter. If GPU load is zero, the machine is not mining productively.
Laptop preflight
Safe to run first. It checks readiness without installing packages or writing miner config.
curl -fsSL https://drinknile.com/install.sh | bash -s -- --preflight --worker 'default'
Install on a supported laptop
Use this only after preflight confirms Linux, NVIDIA/CUDA visibility, and stratum reachability.
curl -fsSL https://drinknile.com/install.sh | bash -s -- --address 'btx1z...YOUR_BTX_ADDRESS...' --worker 'default'
Follow local progress
After starting the miner, watch accepted shares and keep an eye on GPU utilization.
tail -f ~/.dexbtx-miner/miner.log
Check GPU load
For NVIDIA laptops, this should show sustained load while the miner is working.
watch -n 2 nvidia-smi
Local-first rule: if your own laptop can pass preflight and show real GPU work, mine locally. If it cannot, use it to learn the wallet, repository, commands, logs, and expected-yield model before spending on rentals or hardware.
Release-ready setup
No source build, explicit GPU path.
The installer pulls a patched solver binary instead of asking miners to compile BTX source.
The binary hash is pinned and verified before the solver is installed locally.
The installer validates that the GPU is actually doing work instead of silently falling back to CPU.
| GPU class | Expected release path | What the user needs to know |
|---|---|---|
| Pascal, GTX 9/10-series | Native sm_61 cubin | Older cards can run the matmul kernel without building locally. |
| Turing / Ampere, sm_70-sm_86 | sm_61 PTX JIT | First launch may compile the kernel for the installed driver. |
| Ada, RTX 40-series | Native sm_89 cubin | Use the current release for the native Ada binary path. |
| Hopper, H100/H200 | Native sm_90 cubin | Datacenter cards are supported by the prebuilt solver path. |
| Blackwell, RTX 50-series | Native sm_120 cubin | The installer chooses higher worker/thread defaults for faster cards. |
GPU ranking
Find your card and expected BTX/hour.
Estimates use the current network snapshot and the post-trial BTX Start platform fee.
Workers, solver threads, and batch size for modern NVIDIA cards unless a measured row says otherwise.
Used for expected yield. Higher network hashrate lowers BTX/hour; lower hashrate raises it.
BTX targets 90-second blocks, or about 40 blocks per hour before variance.
Expected gross yield for the canonical 28K n/s 5060 Ti reference at the current snapshot.
Yield formula: BTX/hour = gpu_nps / network_nps * 20 * 40.
Measured rows use known-good profile data from local tuning docs; estimated rows are family projections for search and planning.
Your real result is the accepted-share rate your rig sustains.
Loading GPU rows...
| Rank | GPU | Profile | Hashrate | BTX/hour | BTX/day | Efficiency | Confidence |
|---|---|---|---|---|---|---|---|
| Loading GPU ranking... | |||||||
Rent GPUs
Live Vast.ai pricing ranked for BTX mining.
Loading Vast marketplace snapshot...
Verified, rentable Vast offers from the latest marketplace snapshot.
Raw hourly GPU rental price. Vast confirms final cost at checkout.
Estimated from Vast price and the BTX Start GPU yield model.
Expected BTX/hour before pool luck and rental interruptions.
All offer buttons use the configured Vast referral link when active.
BTX Start may earn Vast.ai referral credits or commission from GPU rental links. This does not affect your BTX payout address, pool payout, or mining rewards.
Pricing comes from Vast.ai and is cached on this site for reliability.
| Rank | Vast offer | Price | BTX/hour | Cost / BTX | Region | Reliability | Rental link |
|---|---|---|---|---|---|---|---|
| Loading Vast GPU offers... | |||||||
Track your miner
Know when shares and payouts are moving.
1. Watch your miner log
After launch, accepted shares should appear within a minute or two.
share OK ...
2. Check live stats here
This page shows active workers, recent blocks, network hash, node peers, and the fee currently reported by the backend.
Jump to live stats3. Estimate yield
Use the GPU ranking to compare expected BTX/hour against the current network snapshot.
Open GPU ranking4. Confirm GPU work
Your GPU should show sustained load while mining. Low or zero utilization means the solver is not really using CUDA.
nvidia-smi
Personal view
Signals a miner can watch today.
btx1z...your_address.default
tail -f ~/.dexbtx-miner/miner.log
watch -n 2 nvidia-smi
BTX/hour = gpu_nps / network_nps * 800
Check the wallet that owns your BTX address
Network stats update on this page
Transparency
What is visible, what is first-party.
This page does
- Generate an install command from the visitor's BTX address and worker name.
- Show searchable GPU profiles, expected BTX/hour, and visible yield assumptions.
- Keep installer and source links inside our hosted funnel.
- Show public pool data, payout policy, fee destinations, and backend readiness.
This page does not
- Charge a BTX Start platform fee today.
- Route new miners through the old pool by default.
- Custody funds, collect addresses, or store visitor input.
- Guarantee a user's GPU, driver, network, or backend uptime before the cutover checks pass.
Fee and treasury
- Owned backend fee
- 0.00% first 7d, then 0.50%
- Fee address
Pending first-party backend- Treasury address
Pending first-party backend
First-party backend targets
- InstallerOur GitHub Pages
- Miner sourceOur fork
- Stratum endpointstratum.drinknile.com:3333
- Stats APIapi.drinknile.com/stats
- Per-wallet dashboardBTX Start backend in progress
Public platform treasury
Every platform-fee address is public.
Target platform fee: 0.00% first 7d, then 0.50%
Dedicated backend fee wallet is not connected yet. Once created and deployed, post-trial platform fees should be paid to the public address below.
- Fee address
Pending wallet creation- Treasury address
Pending wallet creation- Total address content
0.00000000 BTX- Balance source
Pending first-party wallet index
Each payout address receives a 7-day 0.00% platform-fee trial. After that, the target platform fee is 0.50% and is intended to fund BTX Start infrastructure, security, miner tooling, and collectively selected new BTX projects.
Treasury spending does not create miner ownership, dividends, or profit-sharing claims.
This will show the fee destination reported by the owned backend once the stats API is live.
- Backend fee address
Pending first-party backend- Backend pending fee
0.00000000 BTX
Fee policy
7 days fee-free, then 0.50%.
Every payout address gets seven days from its first accepted share before the platform fee begins.
The first-party backend should deduct the post-trial fee in payout accounting and route it to the public fee wallet.
The next product milestone is a first-party per-wallet view for shares, worker state, balance, and payouts.
Current BTX Start policy: trial_days = 7,
trial_fee_bps = 0, and
post_trial_fee_bps = 50. The backend should key the
trial by payout address, record the first accepted share, and
deduct the fee from pool payout accounting after the trial ends.
This should be backend-enforced and visible. The miner command remains simple; the pool backend handles fee splits, treasury accumulation, and per-wallet dashboard reporting.
Agent-ready mining
The next operator is software.
The article frames BTX mining as a closed loop an AI agent can operate: install, run, co-schedule, monitor, and settle. The pool client already covers the basic run loop; the next improvements are a machine-readable control surface and idle-cycle scheduling.
Read the agent thesisInstall and verify
Expose repeatable binary, config, CUDA, and pool checks for autonomous setup.
Co-schedule GPU time
Mine only when utilization headroom exists, then yield quickly to inference workloads.
Monitor as JSON
Publish share rate, reject codes, stale work, solve latency, and pool health for agents.
Settle to PQ rails
Validate post-quantum payout addresses and make treasury assumptions explicit.
Network at a glance
Live mining signals.
Loading live stats...
What is BTX?
A Bitcoin Knots fork with matmul proof-of-work.
BTX replaces SHA-256 mining with a deterministic matrix multiplication kernel for commodity CPUs and consumer GPUs, while keeping Bitcoin-like monetary limits and a fast target block time.
90-second target blocks
Difficulty adjusts every block with ASERT, keeping the chain responsive as network hashrate shifts.
21 million coin cap
The current subsidy is 20 BTX, with the first halving planned at block 525,000.
Shielded transactions
BTX includes shielded outputs so amounts, senders, and receivers can be encrypted on-chain.
No premine or ICO
Every circulating coin has been mined, and current liquidity is peer-to-peer or OTC.
Why this pool?
Less node work, more useful hashrate.
No daemon to run
The backend runs btxd continuously. You run the miner and point it at stratum.
No chain to sync
Skip hours of CPU-heavy initial block download before submitting your first share.
Peer mesh maintained
The backend monitors peers, handles stalls, and routes new blocks quickly through the network.
Open source tooling
The client, installer, docs, and release tooling are published in our fork. Binary artifacts still need independent releases.