Guides and news from the Flox and Nix world
Putting the flox and floxify agent skills through their paces with Claude Code: building a FastAPI embedding service from scratch, then reconstructing a complete six-service Mastodon development environment from an existing checkout.
Flox v1.16 introduces flox run, a subcommand that runs commands from Flox Catalog packages without creating an environment, along with pre-built secrets integrations for 1Password, HashiCorp Vault, OpenBao, and Infisical.
Flox v1.14.1 improves activation, search, and usability as more developers, CI systems, and AI agents rely on Flox.
Flox now supports granular, role-based tokens that enable organizations to reduce their blast radius, keep pipelines running smoothly through staff changes, easily rotate credentials on a schedule, and seamlessly authenticate in browserless workflows.
Native Flox auto-activation enables a seamless, friction-free flow by dynamically discovering, activating, layering, and deactivating Flox environments as you move through your projects.
Imagine creating declarative, reproducible environments for the toolchains and dependencies you use in your workflows. Declaring compilers, debuggers, dev libraries, and other dependencies as code. Declarative patterns make this convenient, reproducible, and scalable.
Learn how AI stacks as code make ML and AI environments reproducible across macOS, Linux, Windows, CUDA, and Apple Silicon with reusable reference patterns.
Shai-Hulud-style supply chain worms exploit the secrets, permissions, and network access exposed during builds. Reproducible, hermetic build environments can limit the blast radius across local development and CI/CD.
This article describes a pattern for creating reproducible runtime environments for ML/AI using declared, graph-backed environments based on Nix and Flox. The same Nix and Flox environments run as-is during both model training in local dev to checkpoint validation in eval.
Run project runtimes on your local machine with Nix, Guix, or Flox, let containers handle PostgreSQL, Redis, and other services. Get fast, reproducible local dev that travels from laptop to CI to production ... and back again.
A practical guide to creating and working with Flox standardized local development environments, including playbooks for defining, reproducing, and managing project environments at scale.
This week’s Radio Free Flox dishes on the new packages, environments, AI tools, etc. that land in the Flox Catalog + go live on FloxHub. It also spills tea on what’s going on internally at Flox—including just how much some Flox employees spend on API tokens each month.
CLIs like DeepSeek TUI gives users a fast, terminal-native harness for running local and cloud models on NVIDIA CUDA and Apple Silicon. This article explores how Nix and Flox make it straightforward to run, build, and share the stacks we use to run models locally: model servers, CLI harnesses, and frontier checkpoint models like Qwen 3.6.
Local development is just the first surface of an organization-wide delivery platform. Platform teams own the shared runtime contract that engineers depend on to build, test, debug, and deploy. Flox gives them a way to provide a pinned, declared set of packages, services, and environment variables.
Works-on-my-machine is symptomatic of an ongoing failure to treat the local development environment itself as shared project infrastructure. Standardized development environments are a compelling solution to chronic problems of this kind.
AI is accelerating the rate of CVE discovery. Learn how Flox and Nix transform package vulnerability triage from repeated artifact scans into quick, queryable dependency-graph analysis
Use Flox to run frontier coding models locally on macOS, Linux, or Windows via WSL2 with turn-key GPU-accelerated Ollama and LM Studio environments for Claude Code, Codex, OpenCode, and more.
Use Flox to create declarative, composable ML/AI stacks that travel from Slurm training clusters to Kubernetes production—without rebuilding images or debugging runtime drift.
Test drive OpenClaw and 50+ other agentic AI assistants on your own machine. At once. Without permanently breaking anything. Share what you build with others and it Just Works. Impossible, you say? Not with Flox!
Learn how to pair Flox with containers for local development: Flox provides the project runtime on the host, while containers run supporting services like PostgreSQL.
The BYOA governance stack from Red Hat provides the ceiling and walls. Flox provides the floor: a deterministic, input-addressed runtime substrate that makes the governance layer mean something.
The Flox CUDA Kickstart Program gives teams validated, reproducible CUDA environments and slim GPU-specific builds for common serving workloads, with technical resources demonstrating simple, controlled promotion, one-step rollback, deterministic SBOMs, and other features.
Learn how to use Flox and Nix to build, package, and publish optimized PyTorch packages that are up to 70% smaller than generic builds, with better performance and reduced attack surface.
Learn how to use Flox to define declarative, reproducible model serving environments for NVIDIA Triton, vLLM, and llama.cpp that work across the entire SDLC, from dev laptops to production GPU clusters.
Learn how firms in capital markets use Flox to deliver ML/AI workloads faster and more reliably, with deterministic SBOMs, controlled rollbacks, and the ability to run conflicting dependency stacks side by side.
A look at the Planet Nix 2026 agenda and what Flox is bringing to the conversation around AI, reproducibility, and the evolving SDLC
The Flox VS Code extension integrates cleanly with the industry’s most popular IDE. It makes it simple to create, modify, activate, and monitor Flox environments from directly within your VS Code daily driver.
AI is shrinking the patch window to minutes. Input-addressed hashes + SBOM-driven policy enable you to prevent, detect, and remediate automatically.
In just a few seconds or minutes, an AI coding assistant can take any arbitrary repo and define it as a Flox environment that builds, tests, and runs reproducibly anywhere.
On your own or with an assist from agentic AI, it is shockingly simple to migrate projects into portable, reproducible Flox environments. This article Floxifies two real repos: Node-RED and n8n.
You can use Flox to build, package, and publish with pure Nix expressions. This is extremely useful when you need iron-clad reproducibility guarantees across time and space: what you build and ship locally builds and ships when you push it to CI.
Flox lets you run imageless workloads on Kubernetes using declarative environments in Pod templates for faster, more reproducible deployments.
Kubernetes, Uncontained lets you run imageless workloads on Kubernetes. Workloads start faster, run more reproducibly.
How to migrate Python projects from venv to Flox for reproducible, declarative environments that work the same across every machine, CI, and production.
If you’re vibe coding or prototyping, you can prompt Cursor, CoPilot, Claude Code, Codex, and other tools to quickly and reliably create Flox environments for workflows, and workloads, of all kinds.
We just announced a $25 million Series B funding round led by Addition, with participation from NEA, the D. E. Shaw group, Hetz, and Illuminate Financial. Read about it here.
Get portable, reproducible CUDA environments and stacks with Flox. CUDA stacks you build locally travel as-is across your SDLC. Pull or clone them into runners in CI, then deploy them as containers, VMs, or lambdas to prod.
Teams can now easily build and deploy portable CUDA runtimes—including software like the CUDA Toolkit, Tensorflow, TensorRT, PyTorch, OpenCV, and other packages—that run anywhere, at any time.
Create reproducible Python environments with Flox. Declaratively define everything in one file and activate the same setup every time.
Flox unlocks the potential of Nix without requiring in-depth knowledge of it, and gives Nix experts a great way to enable their teams.
Layering and composition are two powerful ways to create and share Flox environments.
The most important problem in packaging is how to efficiently, reliably, and securely packaging and distributing software so builds are deterministic and runs are portable and reproducible
With Flox Build and Publish, your local dev environment is the same as your CI environment, which are the same as your build and runtime environments. You use the same environment across the entirety of the SDLC.
When a list as influential as the InfraRed 100 includes a company like Flox, we take it as validation of our founding belief: that reproducible runtimes and deterministic builds form the bedrock of scalable, reliable infrastructure—locally, in the cloud, wherever … and whenever.
ComfyUI is a node-based system for creating Stable Diffusion pipelines, and it offers endless possibility. Installing it with Flox is a breeze.
CI/CD is just a wrapper. To align and scale software delivery across the SDLC, teams need environments that are reproducible, auditable, and portable by default.
Flox, the company bringing reproducible dev environments to teams and enterprises everywhere, is thrilled to announce its sponsorship of the Rock Bottom Curling Club.
Every new abstraction we add—containers, microservices, serverless—solves one problem but creates others. This is the story of how I discovered Nix and learned to love a deterministic approach to building software.
You can create a Flox environment with built-in, transparent auth via a third-party secrets manager, giving you CLI workflows that work the same way everywhere—locally, in CI, and in prod.
People often say they don't use Nix because it solves the same problems as containers. This is a half truth. Nix specializes in packaging software, containers excel at deploying it. They're best used together.
Kelsey Hightower will be one of two dozen speakers beaming onto Planet Nix, the annual gathering of Nix users and developers that’s co-located with SCALE in Pasadena, Calif., from March 5th-7th.
Our Flox Ollama runtime environment now includes a popular React-based web front-end, so it looks and behaves like ChatGPT. You can pull and run DeepSeek R1—then prompt it from your browser.
In the real world, Flox environments and containers neatly complement one another. The real magic lies in using Flox inside containers to fetch and manage runtime dependencies at build time.
With Flox, you can easily replicate your preferred apps, tools, settings, and workflows across all your devices—including that brand new M4 MacBook Pro you’re getting for the holidays!
Do your teams need a reproducible, collaborative way to manage AWS secrets and authenticate with AWS services across disparate systems and environments? Fear not! Flox has got them covered.
How to use Flox to install all the runtime dependencies you need to build and test your Lambda functions locally, and also containerize them for deployment to AWS Lambda.
Flox lets you manage both your code and your project environment’s runtime dependencies in the same repository, so your environment’s configuration becomes part of your codebase.
Teams are held back by the unpredictability of things they should be able to control. With shared Flox environments, working on a team becomes a completely different experience.
Onboarding onto something new is radically different when you use Flox to create and manage your dev environments.
Flox’s superpower is reproducibility, so the Go dev environments you create on your Linux laptop will just work ... everywhere
Build consistent Node.js environments across Linux, macOS, and Windows. Flox uses declarative manifests to ensure identical setups for all developers.
By testing your environments as part of your CI workflows, you can guarantee they work the same way—everywhere.
With FloxHub, it’s easy to share and manage Flox environments, making them available to your teammates. Even better, FloxHub provides a one-stop shop for changing your environments.
Flox makes it trivially easy to create and share portable, reproducible Python development environments using simple declarative methods.
See how Flox makes it easy it is to set up a real-world Rust project, complete with external tools and dependencies, using a simple manifest that works everywhere you do.
Flox makes it simple to create portable, cross-platform dev environments, but sometimes you also need to work around system- and architecture-specific package constraints. This article shows you how.
It’s even easier to build beefy services like PostgreSQL into your dev environments now that Flox does the setup, teardown, and runtime management lifting for you.
Flox Package Groups give you an easy way to manage dependency conflicts, enabling you to create powerful fit-for-purpose dev environments you can take with you—anywhere.
Learn how you can use Flox's [hook] and [profile] sections to build, customize, and share rich, reproducible development environments.
Using Flox simplifies Ruby development by making it much easier to manage system-level dependencies, providing a consistent and reproducible Ruby dev environment across all platforms and architectures.
Using Flakes in Flox give you more flexibility, allowing you to access or create unique package versions and custom configurations not available in Flox Catalog.
With the Flox default environment, you can take your $HOME with you wherever you go...and keep your tooling consistent across systems and distros.
Using Flox, it is easy to create reproducible environments for ML training and inference using Hugging Face Diffusers. Our FLAIM environment gets you there in a single command.
Flox environments aren't isolated; they layer upon one another, allowing you to combine them in endless ways while keeping your toolsets discrete.
The Flox Catalog now contains over 3 years of history for each of its 100k packages, totaling over a million building blocks you can assemble however you like. Find everything you're looking for.
The third part of this series shows how you can clone environments, layer them, and share them.
The second part of this series explores the ways you can use Flox environments to adjust the way software is configured and run.
Flox makes it easy to create ephemeral environments and try new tools without making a mess all over your desk.
Flox CLI 1.0 and FloxHub is here! It's a good opportunity to reflect on our own experiences and the reasons we feel Flox is a part of the next generation of software development practices.
Nixcon 2023 in Darmstadt is a wrap! The team from flox was there, giving talks and having amazing conversations. In this post, find videos from our NixCon 2023 talks.
If Rust seeks to solve the problem of memory management at the program level, Nix seeks to solve dependency management at the build level, for any language. The solutions go hand in hand.
We're big fans of Nixpkgs (obviously) and encourage others to contribute packaging and review. Nixpkgs is surprisingly easy to contribute to, but has its particular details. This is a guide to contributing your first PR to Nixpkgs.
A clear guide to Nixpkgs, the world’s largest open-source package collection. Learn how its recipes enable cross-platform builds and community curation.
Harness the power of Nix. Today we’re thrilled to release the flox CLI beta to the public as an open source project and announce a Series A funding round led by NEA.
Github changed the format of their archives and this caused some systems and services that depend on the hash of the archives of a source tree to break.
Tired: year of the Linux desktop. Wired: year of Nix everywhere.
At NixCon 2022, Flox Lead Engineer Tom Bereknyei talked about the mission behind Flox, focusing on the problems we can solve ... and how we intend to go about solving them.
Last week, the Nix community got together in person for the first time in 3 years, in Paris, for three days of presentations, discussion, and good times.
flox is a first-of-its-kind environment manager. Seamlessly create new polyglot developer environments. Make everything reproducible from package management to builds.
I absolutely love it when hard problems are solved by challenging convention, peeling back layer after layer until pinpointing the core assumption that led us down the wrong path in the first place.
