Get Started

Execute work. Not replicas.

PitBox prepares portable Components and schedules each execution directly across a shared pool of execution lanes.

QUEUE:
8 Shared Lanes
L01
service-a
L02
service-c
L03free
idle
L04
service-b
L05
service-a
L06free
idle
L07
service-d
L08
service-b
8
Execution lanes
Configurable bounded concurrency per host
7+
Source languages
All executing on unified Wasmtime sandboxes
0
Per-service ports
Demand-driven execution on free slots
EXECUTION PARADIGM

High-density shared compute without containers

Traditional container orchestrators hold idle memory for thousands of static replicas. PitBox only schedules active work.

01Capacity

Shared Execution Grid

All services compete for the same available execution capacity. When a request ends, the lane is freed instantly.

02Wasmtime

Prepared Components

Compile once, reuse prepared Wasmtime artifacts across executions for sub-millisecond execution start times.

03Isolation

Bounded Scheduling

Queue depth and lane capacity keep concurrency strictly controlled without reserving static capacity per service.

EXECUTION BENCHMARK

Run and benchmark workloads directly

Execute tasks locally or fire high-throughput load tests through the PitBox execution lanes.

bash
# 1. Execute workload directly inside a lane
pit run service-a:v1

# 2. Run high-concurrency benchmark against shared lanes
pit bench pit://service-a/hello --concurrency 50 --requests 5000

# 3. View real-time lane utilization and queue depth
pit status
The PitFast Lifecycle

How it connects to the rest of PitFast

Every subsystem handles one distinct responsibility in the demand-driven execution chain.

RUN WITHOUT OVERHEAD

Put the Grid to work.

Start the PitFast daemon locally and watch requests execute across disposable lanes.