🪐 Pluto Launches: OSS Experiment Tracker for Neptune Users

Pluto by Trainy recently launched!

^‍

^TL;DR:

^{→ Try the live playground – explore Pluto with real sample experiments, no signup required.}

^{NeptuneAI is shutting down on March 5th after being acquired by OpenAI. They built Pluto so you can migrate safely without rushing.}

‍

Founded by Roanak Baviskar & Andrew Aikawa

The pair founded Trainy to help AI teams efficiently manage and utilize GPU infrastructure. When they heard Neptune was shutting down, their customers started panicking about losing years of experiment history and having to rewrite their logging code on a tight deadline.

So they built Pluto.

‍

🔥 The problem

Neptune's March 5th shutdown is forcing research teams into one of the worst kinds of migrations: rushed, high-stakes, and hard to verify.

Most alternatives require you to:

• Rewrite your logging code
• Lose your historical runs (or spend days exporting and reformatting)
• Hope the new tool handles your multi-GPU workloads the same way

And you have to do all of this before March 5th or your data gets deleted.

‍

🛠 How Pluto solves it

1. Dual-log to both platforms simultaneously

Add one import, keep your existing Neptune code running, and Pluto logs side-by-side. Validate that everything matches on your real training workloads before you cut over.

2. Export your Neptune history

Pluto includes a Neptune exporter that brings your old runs into Pluto. Years of experiment data, preserved.

3. Built for the workloads Neptune users actually run

They are optimizing for what Neptune users loved: UI responsiveness at scale, reliable logging throughput under multi-node/multi-GPU workloads, and the ability to track thousands of per-layer metrics without lag.

Deployment options:

• Self-hosted – for sensitive data or air-gapped environments, instructions in server readme
• Hosted – $250/seat/month (free for existing Trainy customers)

‍

🙏 The ask

If you're a Neptune user, please try Pluto with dual-logging before the March shutdown.

They are specifically looking for feedback on:

• Anything that would block your cutover
• Features you rely on that they are missing
• UI/performance differences at scale

Early Bird offer: Start dual-logging in the next 2 weeks → 3 months hosted free.

‍

🔗 Links

• Playground – try it now
• Quickstart – get running in 5 minutes
• GitHub – Apache-2.0 (client), AGPL-3.0 (server) ⭐

They are listed on Neptune's official transition hub.

‍

Learn More

^‍

^{🌐 Visit www.trainy.ai to learn more.}

^‍

^{👉 Make an account on pluto.trainy.ai, email Roanak here or grab a time to chat and he will add seats/storage.}

^‍

^{🤝 Playground – try it now}

^‍

^{⭐ GitHub – Apache-2.0 (client), AGPL-3.0 (server) ⭐}

^‍

^{👣 Follow Trainy on LinkedIn & X.}

‍