frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Marimo pair – Reactive Python notebooks as environments for agents

https://github.com/marimo-team/marimo-pair
16•manzt•1h ago
Hi HN! We're excited to share marimo pair [1] [2], a toolkit that drops AI agents into a running marimo notebook [3] session. This lets agents use marimo as working memory and a reactive Python runtime, while also making it easy for humans and agents to collaborate on computational research and data work.

GitHub repo: https://github.com/marimo-team/marimo-pair

Demo: https://www.youtube.com/watch?v=6uaqtchDnoc

marimo pair is implemented as an agent skill. Connect your agent of choice to a running notebook with:

/marimo-pair pair with me on my_notebook.py

The agent can do anything a human can do with marimo and more. For example, it can obtain feedback by running code in an ephemeral scratchpad (inspect variables, run code against the program state, read outputs). If it wants to persist state, the agent can add cells, delete them, and install packages (marimo records these actions in the associated notebook, which is just a Python file). The agent can even manipulate marimo's user interface — for fun, try asking your agent to greet you from within a pair session.

The agent effects all actions by running Python code in the marimo kernel. Under the hood, the marimo pair skill explains how to discover and create marimo sessions, and how to control them using a semi-private interface we call code mode.

Code mode lets models treat marimo as a REPL that extends their context windows, similar to recursive language models (RLMs). But unlike traditional REPLs, the marimo "REPL" incrementally builds a reproducible Python program, because marimo notebooks are dataflow graphs with well-defined execution semantics. As it uses code mode, the agent is kept on track by marimo's guardrails, which include the elimination of hidden state: run a cell and dependent cells are run automatically, delete a cell and its variables are scrubbed from memory.

By giving models full control over a stateful reactive programming environment, rather than a collection of ephemeral scripts, marimo pair makes agents active participants in research and data work. In our early experimentation [4], we've found that marimo pair accelerates data exploration, makes it easy to steer agents while testing research hypotheses, and can serve as a backend for RLMs, yielding a notebook as an executable trace of how the model answered a query. We even use marimo pair to find and fix bugs in itself and marimo [5]. In these examples the notebook is not only a computational substrate but also a canvas for collaboration between humans and agents, and an executable, literate artifact comprised of prose, code, and visuals.

marimo pair is early and experimental. We would love your thoughts.

[1] https://github.com/marimo-team/marimo-pair

[2] https://marimo.io/blog/marimo-pair

[3] https://github.com/marimo-team/marimo

[4] https://www.youtube.com/watch?v=VKvjPJeNRPk

[5] https://github.com/manzt/dotfiles/blob/main/.claude/skills/m...

Comments

manzt•4m ago
One of the authors here, happy to answer questions.

Building pair has been a different kind of engineering for me. Code mode is not a versioned API. Its consumer is a model, not a program. The contract is between a runtime and something that reads docs and reasons about what it finds.

We've changed the surface several times without migrating the skill. The model picks up new instructions and discovers its capabilities within a session, and figures out the rest.

Hot-reload for your database schema with MooseStack

https://github.com/514-labs/moosestack
1•satyampsoni•38s ago•1 comments

Anthropic Claims Its New A.I. Model, Mythos, Is a Cybersecurity ‘Reckoning’

https://www.nytimes.com/2026/04/07/technology/anthropic-claims-its-new-ai-model-mythos-is-a-cyber...
1•jbegley•1m ago•0 comments

History of Lab Notebooks

https://www.asimov.press/p/lab-notebooks
1•mailyk•1m ago•0 comments

Cameron's World

https://www.cameronsworld.net
1•thunderbong•2m ago•0 comments

The Cognitive Costs of AI

https://garden.johanneskleske.com/the-cognitive-costs-of-ai
1•speckx•2m ago•0 comments

Reasons to be pessimistic (and optimistic) on the future of biosecurity

https://www.owlposting.com/p/reasons-to-be-pessimistic-and-optimistic
1•_ihaque•2m ago•0 comments

This Spillway Failed on Purpose

https://practical.engineering/blog/2026/4/7/this-spillway-failed-on-purpose
1•crescit_eundo•3m ago•0 comments

I built a rejection therapy practice tracker

https://bubbling.dev/practice/rejection-therapy
1•georgesmith9914•3m ago•0 comments

The nifty Japanese printing gadget uniting artists worldwide

https://www.theguardian.com/artanddesign/2026/apr/07/nifty-japanese-printing-gadget-risograph-ris...
1•anfractuosity•4m ago•0 comments

Show HN: Gemma 4 Multimodal Fine-Tuner for Apple Silicon

https://github.com/mattmireles/gemma-tuner-multimodal
1•MediaSquirrel•4m ago•0 comments

Show HN: Kronaxis Router – Don't pay frontier prices when a local LLM is enough

https://github.com/Kronaxis/kronaxis-router
1•JasonDuke•5m ago•0 comments

S3 Files

https://aws.amazon.com/blogs/aws/launching-s3-files-making-s3-buckets-accessible-as-file-systems/
3•akshaysaxena•5m ago•0 comments

What AI is Doing to the Workforce [video]

https://www.youtube.com/watch?v=ncmuXQGGqBM
1•johnath•6m ago•0 comments

Where are we converging with all AI companies running the same business model

2•elmlabs•7m ago•0 comments

Show HN: Hive – full dev workspace using (Kanban/chat mode,multi-repo,agent-SDK)

https://github.com/morapelker/hive
1•moropex•7m ago•0 comments

Eyes on the Far Side of the Moon

https://www.nytimes.com/interactive/2026/04/07/science/space/moon-photos-artemis-2-nasa.html
2•saikatsg•8m ago•0 comments

The V32 numbers station: a mysterious Cold War spying system revived in Iran

https://english.elpais.com/international/2026-03-19/the-v32-numbers-station-a-mysterious-cold-war...
2•mazokum•8m ago•0 comments

SCOTUS overturns 5th Circuit ruling that told ISP to kick pirates off Internet

https://arstechnica.com/tech-policy/2026/04/scotus-overturns-5th-circuit-ruling-that-told-isp-to-...
3•CharlesW•10m ago•0 comments

Ask HN: What is the most useful website you've discovered?

1•hopefully_can•11m ago•0 comments

Give Your AI Eyes: Introducing Chrome DevTools MCP

https://addyosmani.com/blog/devtools-mcp/
1•fagnerbrack•12m ago•0 comments

Will you run one reproducibility check? (Voynich structural analysis)

https://github.com/Frederick-Stalnecker/voynich-section-grammarInstructions:REPRODUCE.md
1•TheosResearch•13m ago•2 comments

Show HN: I built an AI coding agent 50% cheaper than Claude Code (same prompts)

https://github.com/kirby88/vix-releases
1•kirby88•13m ago•0 comments

Duolingo-style GitHub streak widget for READMEs

https://github-streak.rahuldhole.com/
2•rahuldhole•13m ago•0 comments

Before We Start on Quantum

https://scottaaronson.blog/?p=9668
1•whatisabcdefgh•15m ago•0 comments

1•illegalmemory•17m ago

Show HN: Visual DLQ monitoring and replay for RabbitMQ

https://queueforgehq.com
1•rootN•18m ago•0 comments

Why IPv6 is the only way forward

https://ankshilp.in/posts/for-the-love-of-internet/
8•quaintdev•19m ago•1 comments

Roger Fenton's Valley of the Shadow of Death (1855)

https://publicdomainreview.org/collection/roger-fenton-valley-of-the-shadow-of-death/
1•speckx•19m ago•0 comments

Apple Studio Display XDR Now FDA-Cleared for Diagnostic Radiology Use

https://www.macrumors.com/2026/04/07/studio-display-xdr-fda-clearance/
2•tosh•20m ago•0 comments

Spacebot: Agentic AI system where LLM process has a dedicated role, OpenClaw alt

https://spacebot.sh/
1•maxloh•20m ago•0 comments