Projects

Python-native systems PyPV builds

Each project is a control plane for a different layer of Python infrastructure: background execution, network management, and the handoff between autonomous agents and human decisions. Built with FastAPI and Django, self-hostable, and operationally honest.

z4j In development

Z4J

Control plane for Python background jobs.

A control plane for Python background work that unifies tasks, workers, queues, schedulers, retries, and observability into one operational layer instead of a pile of disconnected scripts.

The problem. Python teams stitch together Celery, RQ, cron, and one-off scripts with almost no shared visibility. When a job stalls or silently fails, nobody can see why. Z4J gives operators a single place to run, inspect, and control background execution.

Priority task queues backed by Postgres and Redis
Cron and interval scheduling with timezone awareness
Automatic retries with configurable backoff and dead-lettering
Live worker, queue, and throughput visibility
Typed FastAPI control API for every operation

Python
FastAPI
Django
PostgreSQL
Redis
Workers
Schedulers
Retries
Observability

Visit Z4J GitHub

freesdn Open source

FreeSDN

Open-source network management.

A free, open-source, self-hosted network management platform that unifies infrastructure devices and vendors behind one consistent control plane and API.

The problem. Network state is fragmented across vendor-specific consoles, CLIs, and tooling. FreeSDN provides a single self-hosted layer for managing heterogeneous network infrastructure without locking operators into a proprietary cloud.

Vendor-neutral device and interface models
Declarative, version-controllable network state
Self-hosted deployment with no cloud dependency
REST and Python APIs for automation
Unified inventory across multiple vendors

Python
FastAPI
Django
Multi-vendor
Network automation
Self-hosted
Open source

Visit FreeSDN GitHub

ony Early access

ony.ai

Voice handoff for AI agents.

A voice handoff layer for AI agents. When a coding agent (Claude Code, Codex, or a long-running workflow) hits a decision that needs a human, Ony calls the user, captures the answer, and routes it back into the workflow.

The problem. Autonomous agents stall the moment they reach a decision only a human can make. Ony bridges that gap with a real-time voice handoff, so long-running coding and automation workflows keep moving instead of waiting on a missed message.

Outbound voice calls when an agent needs a decision
Structured capture of the human's answer
Webhook and SDK callbacks back into the workflow
Drop-in integration for coding and automation agents
Full audit trail of every approval and decision

Python
FastAPI
Django
Real-time voice
Agent integration
Human-in-the-loop
Approvals

Visit ony.ai GitHub

Note: the GitHub organization links above are placeholders pointing at github.com/pypv until each repository is confirmed public.