maniva
ENESFRPT

Share one link. Get the evidence, protocols, data and analysis behind the research.

Agricultural research leaves papers, protocols, data and code in separate places, and using them takes someone who already knows where to look. maniva links and documents them, so anyone, and any AI assistant, can use them without that insider.

tricot is the first domain, with on-farm trial data on more than 30 crops in more than 25 countries since 2011. Weather, elevation and soil data are available for any site, and household data come next.

Four layers, one entry point

Each layer lives in its own repository, with its own version, licence and DOI. maniva routes each question to the layer that answers it. Pick a question to see the route.

Questions people ask
  1. Knowledge base

    What does the published evidence say?

    Verified cards on the open-access tricot literature. Findings carry their primary citation and the people behind them.

    Active, assistant guide ready

  2. Protocols

    How do I run a trial?

    Standard operating procedures, surveys and trait metadata in five languages, with traits mapped to the Crop Ontology.

    Available, assistant guide planned

  3. Data

    What did the trials find?

    Harmonized trial data in JSON, one file per trial, archived on Zenodo under CC0, plus daily weather, elevation and soil for any site.

    Available, assistant guide planned

  4. Analysis

    How do we reason with the data?

    Curated use cases with decision records on methods and their limits, the R packages they run on, and a template for reproducible, FAIR projects.

    Active, assistant guide ready, packages on CRAN

Use it in three steps

Open an AI assistant that can read the web
Works best with assistants that can open several files from GitHub in one conversation, such as Claude with web access. Nothing to install, no account beyond your assistant.
Paste the maniva link with your question
Share github.com/AgrDataSci/maniva and ask in the same message. The assistant finds the layer that answers it.
Check the source it cites
Every finding should name the paper, protocol or dataset behind it.

Read the full guide

How we keep it trustworthy

Collaborate

maniva grows through the people who use it. Each layer is a paper, and partners who shape a layer join it as co-authors.

Talk to the team

Write to any of us about data, partnerships or funding. All three of us are at the Alliance of Bioversity International and CIAT.

Why maniva

Cassava is propagated from the maniva, a stem cutting that carries the potential for new growth. The crop was domesticated in the Amazon and travelled to Africa and Asia, adapting to new environments on the way.

maniva is meant to work the same way, so that knowledge developed in one place can take root, adapt and create value elsewhere.

Rooted in biodiversity. Growing through collaboration.

Cite maniva

When you use a specific layer, cite that repository too, and cite the original papers for any finding.

de Sousa K, Laporte MA, Youngberg B, Cremaschi A, Chase R, van Etten J (2026). maniva. An open commons of evidence, protocols, data and analysis for agricultural research. Version 1.0.0. https://doi.org/10.5281/zenodo.24011051

Machine-readable metadata is in CITATION.cff.