GoKawiilTithon lets Jupyter kernels keep running after VSCode disconnects
A new open-source tool called Tithon decouples Jupyter kernel execution from the client, running the kernel and streaming its output through a host-side daemon instead of the editor process. This means closing VSCode, losing SSH, or shutting a laptop no longer kills a long-running notebook session, and reconnecting restores the full output history rather than a blank slate. It also stores notebooks as compact .py files with real image outputs instead of bloated .ipynb JSON, making them easier for AI agents to read.
GoKawiil's interpretation of the reporting above, not reported fact.
Data scientists and ML engineers running long GPU jobs have long suffered from session loss caused by client-tied kernels in tools like JupyterLab and VSCode Jupyter, forcing workarounds like tmux that sacrifice rich output. By moving session state to the host, Tithon addresses a structural gap in how Jupyter handles persistence and reconnection. Its cleaner file format also targets a growing use case: AI coding agents that need to parse notebook outputs efficiently without wasting tokens on base64-encoded JSON.
- Tithon runs a persistent daemon on the host so Jupyter kernels survive client disconnects or crashes.
- It stores notebooks as lean .py files with real image files instead of verbose .ipynb JSON, aiding AI agent workflows.
- The project is in alpha, requires Python 3.11+ and a Unix-like host, and is in active daily use by its developer.
Anker 737 Power Bank (PowerCore 24K) — If you're the kind of person babysitting long-running notebook jobs over SSH from a laptop, keeping that laptop alive matters as much as keeping the kernel alive. The Anker 737 is a beefy USB-C power bank that can top up a laptop and phone on the go, so you can close the lid and move without hunting for an outlet. A solid companion for persistent remote sessions.
See Anker 737 Power Bank (PowerCore 24K) on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: github.com, 2026-09-09
Published there as: “Show HN: Persistent Jupyter kernel execution and live output streaming in VSCode”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.