Checkpointing Python Packages

sayiir

Sayiir: simple, embeddable durable workflow engine in Rust with Node.js & Python bindings. Checkpoint-based recovery, no deterministic replay. A simplified alternative to Temporal, Restate, and Airflow.

sylo-sdk

Sylo is an open-source Python SDK and cloud platform that adds production-grade reliability to your existing AI agent pipelines — checkpointing, permission enforcement, human approval gates, and immutable audit logging — without replacing the frameworks you already use.

chidori

The agent framework where every run is durable, replayable, and resumable by default.

dlio-benchmark

An I/O benchmark for deep Learning applications

dagpipe-core

Zero-cost, crash-proof LLM pipeline orchestrator. Features disk-based checkpointing, free-tier routing, and structured output. (LangGraph / CrewAI alternative)

torchgpipe

A GPipe implementation in PyTorch

saveable-objects

A Python package for checkpointing, saving, and loading objects.

thread-chunks

A Python package for performing memory-intensive computations in parallel using chunks and checkpointing.

keras-buoy

Keras wrapper that autosaves what ModelCheckpoint cannot.

hyperphoenixcv

Hyper-parameter search that rises from the ashes of interrupted experiments.

agentstateprotocol

Open-source checkpoint protocol for regulated AI. Replay, audit, branch, and recover every AI decision with full reasoning traceability.

agent-recall-ai

Structured session checkpointing for AI agents — survive context limits, cost overruns, and session death

checkpnt

Framework-agnostic state persistence for AI agents. Resume any agent from exactly where it stopped.

wandb-preempt

Code and tutorial on integrating wandb sweeps with Slurm pre-emption