RPC Gateway Logic Engine
The core of rpc.ag is the Rust-based Load Balancer. It abstracts the complexity of managing multiple RPC providers (Helius, Triton, QuickNode, etc.). Below is an interactive simulation of how our "Method-Based Routing" functions compared to standard Failover strategies.
⚙️ Configuration
Requests are distributed sequentially (A -> B -> C). Good for cost saving.
Live Stats
Deep Observability
We prioritize giving developers total visibility. We use Clickhouse to store high-cardinality data for instant analytics on every single request.
- ✓Full Request Logging: Capture bodies, headers, and responses.
- ✓Security Fingerprinting: Log IP addresses and TLS fingerprints to block abuse.
- ✓Performance Metrics: Breakdown latency by provider, method, and region.
Why Clickhouse?
Postgres is great for user data (Accounts, Keys), but Clickhouse is essential for the sheer volume of logs (Millions of RPC calls) required for real-time analytics without slowing down the gateway.
Global Request Latency (ms)
The Architecture Stack
Rust Server
The "Gateway". Handles incoming traffic, TLS termination, and strategy execution.
- • Memory safe & blazing fast
- • Async Tokio runtime
- • Custom Load Balancing logic
PostgreSQL
The "Source of Truth". Stores user configurations, API keys, and persistent state.
- • Relational data integrity
- • User Accounts & Projects
- • Provider configurations
Clickhouse
The "Analyst". Ingests massive streams of logs for real-time querying.
- • Columnar storage for speed
- • Peta-byte scale logging
- • Instant aggregate queries
Developer Experience First
We aren't just building infrastructure; we are building tools.
Integrated Playground
Build, save, and share RPC requests. Export to cURL or TypeScript instantly. No more boilerplate.
Devnet Support
Test without fear. Full support for Devnet clusters with integrated faucets, allowing teams to simulate production environments for free.
Latency: 45ms
Provider: Helius (Fastest)