Bare-metal infrastructure. Honest pricing.
No seat fees. No compute units. No overages. Flat monthly rate, unlimited requests.
── latency tester ──
va-bsc-01 · ashburn, va30 quick round-trips from this device to our BSC node. Mobile / wifi will see your network last-mile latency. Your bot running on a VPS will see lower.
Quick check via your browser — adds ~1–3 ms of JS overhead and can't separate network from server-side latency. For the SLA-grade measurement, run our CLI: streamsuite-bench →
BNB Chain endpoints
Bare-metal BSC infrastructure. Each tier unlocks additional methods and support levels. Latency shown is server-side processing time. Your total round-trip depends on your network distance to Ashburn, VA. Active colocation group with open slots. Typical provisioning in under an hour.
Every tier is unmetered. No rate limits, no compute units, no overages. The number you see is the number you pay. Month-to-month, cancel anytime.
Real-Time
HTTP + WebSocket. No limits.
- HTTP + WebSocket endpoints
- All standard methods (eth_*, net_*, web3_*)
- eth_call, eth_estimateGas, eth_sendRawTransaction
- newHeads and logs subscriptions
- ~4.5 days block, log & receipt history (last 900k blocks)
- 10 concurrent connections (HTTP + WS combined)
- Email support
Mempool
Pending tx stream. For searchers.
- Everything in Real-Time
- txpool_content, txpool_inspect, txpool_status
- newPendingTransactions WebSocket subscription (hashes or full tx objects)
- 20 concurrent connections (HTTP + WS combined)
- Direct Telegram / Discord support (contact sent in welcome email)
- Priority response < 4 business hours
Full Node
Trace access. Priority delivery.
- Everything in Mempool
- debug_traceTransaction on the last ~128 blocks (~60 sec, chain-tip only)
- Sole debug-access tenant: only 1 Full Node customer per server
- 30 concurrent connections (HTTP + WS combined)
- Need deeper trace history? Ask about the Archive add-on
- Direct Telegram to operator (contact sent in welcome email)
- Hardware-level diagnostics on request
Why Full Node is capped at 1 per server.
Two concurrent debug_traceTransaction tenants contend on geth's state trie hard enough to introduce ~hundreds of ms of cross-talk on every RPC method — we benchmarked it. So the Full Node tier gets the entire debug surface to itself, and the price reflects single-tenant economics: pays for the box and still leaves room for the 9 lighter operators to share the rest.
“What if I go over some limit?”
No rate limits, no request caps, no compute-unit math. Fire as many requests per second as your bot can handle — we don't meter volume, we don't bill overages, and you won't get throttled.
The one cap that exists is on simultaneous connections per API key (10 / 20 / 30 depending on tier) — pure anti-abuse so one bad actor can't hog the node. A normal single-bot setup uses 5–11 connections, well under every cap. If you genuinely need more concurrent connections than the Full Node tier provides, ask about a dedicated server (unlisted tier, contact us).
Colocation groups, explained.
We use the term “colocation group” deliberately. It means a capped cohort of operators physically colocated on the same bare-metal server. Not a shared multi-tenant cloud instance.
Each group has its own designated hostname (e.g., va-bsc-01.streamsuite.io), a known capacity ceiling, and measured performance characteristics under full load. Customers know which group they're on.
When a group fills, we bring up a new one rather than squeezing more operators onto existing hardware. Dedicated (single-tenant) hardware is available separately for Custom Builds and colocated bot hosting.
Dedicated node, your chain, 14 days.
Solana, Ethereum, Base, Arbitrum, Optimism, or any EVM chain. We provision dedicated hardware and give you the endpoints. Same iron as BSC. Bare-metal, NVMe, Ashburn.
- Solana
- Ethereum
- Base
- Arbitrum
- Optimism
- Any EVM, ask
We build it. You run it.
From MEV research tooling to production trading bots. Tell us what you need. we scope it, build it, and optionally run it on our hardware with sub-millisecond IPC to the node. No network hops between your bot and the chain.
Custom analysis
Mempool analytics, MEV opportunity scanners, token flow tracers, on-chain forensics, backtest frameworks. Built to your spec against live or historical chain data.
- Mempool pattern detection
- MEV opportunity analysis
- Token flow & wallet tracing
- Historical backtest engines
Custom bots
Arbitrage, liquidation, sniping, market making. Whatever the strategy. Built in Rust or TypeScript, optimized for your edge, delivered production-ready.
- Strategy-specific architecture
- Rust or TypeScript
- Telegram/Discord controls
- PnL tracking & dashboards
Run on our iron
Deploy your bot in an isolated Docker container on the same physical machine as the RPC. Unix socket IPC. Sub-millisecond round trip, benchmarked under full operator load. Faster than HTTP localhost. 3 IPC slots per colocation group.
- sub-millisecond IPC to the node
- Dedicated CPU cores & RAM
- Your keys stay in your container
- Start, stop, adjust via Telegram
Card now. Crypto next.
Card billing live today — Visa, Mastercard, Amex via Stripe, automatic monthly renewal, cancel anytime. Crypto checkout (one-time 30-day windows on major EVM chains) is in final testing — join the crypto waitlist and we'll email when it goes live (target: next 2 weeks).
Still deciding? Read the docs or email us directly.