Bioinformatics has a reproducibility problem. Pipelines live in repositories, runs live in logs, and the connection between a result and how it was generated is frequently lost. Adagio was built to fix that from the ground up.
Workflow languages are powerful, but they ask every user to be an engineer. Running a pipeline means cloning a repository, reading through source code to find required parameters, managing environment configurations, and hoping nothing has changed since the documentation was written.
Building a new pipeline means learning a domain-specific language, reasoning about data flow and execution logic, and accepting that a pipeline that looks correct might silently produce wrong results.
And once a result exists, its provenance begins to decay immediately, living in a log file, tied to a specific run, on a specific machine, that may not exist next year.
Portability
An Adagio pipeline is a single .adg file.
Share it like a spreadsheet. Recipients need one command to run it. No repo
to clone, no dependencies to untangle, no config to decipher. The CLI
exposes parameters dynamically so users interact with a pipeline as though
it were a standalone tool.
Type Safety
Pipeline construction in Adagio happens through a visual canvas where every node has strictly typed inputs and outputs. Incompatible connections are structurally impossible. If the canvas lets you connect two nodes, that connection is correct. Complex pipelines (variant calling, RNA-seq, metagenomics) can be constructed in minutes with confidence they will run.
Provenance
Every Adagio output carries its complete history bundled inside: the pipeline that generated it, every input file, every parameter value, and every package and environment version. This isn't a log you maintain separately. It's baked into the artifact. Drag any result into the Adagio frontend months or years later and see exactly what produced it.
For Your Team
Adagio is designed for three distinct roles that rarely need to interact. Scientists get a simple web form or a one-command CLI. Bioinformaticians get the visual canvas to construct new pipelines without writing code. Engineers get a plugin framework focused on functionality, where a small amount of boilerplate integrates any command-line tool into the ecosystem. Each person only needs to understand their layer.
Adagio is built on top of QIIME 2, a framework with over ten years of development and thousands of citations across the scientific literature. QIIME 2 provides the sophisticated type system and plugin architecture that makes Adagio's safety guarantees possible, which is the reason incompatible connections are structurally impossible rather than just warned about.
Today, Adagio supports diverse microbiome amplicon analyses. The roadmap includes shotgun metagenomics and, over time, a full ecosystem of data science workflows built by Adagio's team or by yours.
High-quality research software is critical for reliable and efficient science, and that's what we build.
At its core, Cymis Benefit Corporation's goal is to promote advancements and opportunities in scientific computing, with an emphasis on reliability, transparency, reproducibility, open source software, and increasing diversity and inclusion in the field. When you use our software, you're supporting these goals while advancing your own.
Based on over four decades of collective experience engineering, supporting, and applying the most widely used microbiome data science software for academia and industry, we've designed the bioinformatics infrastructure that will help you get your analysis done. We've worked with our users to understand their needs, and we built Adagio to address those needs.
Browse community pipelines or join the waitlist to get early access.