Dual engine architecture
A runtime for
Every workload.
Wasmtime 40 for production. Wasmer 7.1 for edge. Routed per workload, not per cluster.
2
Runtimes
<1ms
Cold start p99
0.2
CPU seconds min
15
Regions
Runtime selection
One platform. Different execution paths.
The workload interface selects its execution path. The deployment surface remains consistent across runtimes.
Example runtime classifier
WASI component
wasi:http/proxyInspect interface
WasmtimeComponent-model execution Selected example path
Wasmer / Edge.jsCompatibility executionAlternate runtime
Same deployment surface, selected runtime.
Dual engine.
Two world-class runtimes. One platform. The right engine for every workload.
Production
Wasmtime 40
The Bytecode Alliance reference runtime. Full WASI 0.2 Component Model. Ahead-of-time compilation. Snapshot restore for instant cold starts.
WASI 0.2
AOT Compile
InstaBoot
Component Model
WASI 0.3 Async
Edge
Wasmer 7.1
Lightweight interpreter optimized for edge locations. 200KB runtime binary. WASIX sockets. Perfect for latency-sensitive workloads close to users.
WASIX
Edge.js
Proxy-Wasm
Node.js APIs
200KB Binary
Intelligent dispatch
Automatic routing.
Workload
.wasm
Any language, any size
Runtime selector
Analyze
WASI version, runner type
Wasmtime
Production workloads
Wasmer
Edge + WASIX workloads
You never choose the engine. You describe the workload. Hives inspects the WASI version, runner type, and target environment, then routes to the optimal runtime automatically.
Performance.
Numbers that matter. Measured in production, not in marketing decks.
0.8ms
Cold start
Snapshot restore, not instantiation. Sub-millisecond p99.
50Krps
Throughput
Per-node request throughput. Linear horizontal scaling.
2MB
Memory floor
Minimum memory per instance. Scale to zero when idle.
15
Regions
Global deployment. Workloads run close to your users.
Deploy your first workload.
One command. Sub-millisecond cold starts. Free tier included.