Open-source Python runtime

Portable Python
execution state.

Pause supported Python work, save its live state, and continue in a new process without replaying from the beginning.

Explicit logical framesPortable .cont imagesPublic proof
continuum.runtimeCAPTURE
SOURCE PROCESS
workload()pc 42
middle_layer()pc 18
inner_step()pc 74
PORTABLE IMAGE
.contstructured state
TARGET PROCESS
workload()pc 42
middle_layer()pc 18
inner_step()pc 74
Runtime-owned state Cross-process continuation Inspectable evidence MIT licensed
The idea

A running program should be able to pause without forgetting where it was.

Continuum moves execution state out of an opaque native stack and into a logical model the runtime can inspect, serialize and restore.

How it works

Three steps. One execution.

01

Freeze

Stop at a logical instruction boundary and record the state Continuum owns.

02

Transfer

Package frames, values and supported resources into a validated image.

03

Resume

Load the saved logical state directly in a compatible new process.

Architecture

Portable because the runtime owns the state.

Continuum does not copy a native CPython stack. Supported code runs through explicit logical frames with stable, inspectable fields.

Read the architecture notes ↗
logical_frameruntime-owned
function_id
workload
program_counter
42
locals
object graph
operand_stack
1 item
control_state
resumable
Live from GitHub

The site follows the repository.

Repository activity and CI status are loaded live. No release number or platform claim is hard-coded here.

Checking repository

Recent workflow runs

Open Actions ↗
Loading public GitHub data…
Honest scope

A controlled Python runtime—not arbitrary live CPython migration.

Continuum supports a defined Python subset through its own execution model. The repository remains the source of truth as capabilities change.

OPEN SOURCE
MIT

Use, study and extend the runtime.

Read license ↗
Portable Python execution state.

Pause. Move. Continue.

Explore Continuum ↗