BUILDING IN THE OPEN

A useful commons.
A clear way to help.

Open code becomes useful when people can understand it, test it and improve it together. We are building those foundations.

Choose a small first contribution

1. Test a workflow. Try the example, save labels and reopen them. Report your device, steps, expected result and actual result. Keep recordings and personal details private.
2. Improve one evidence note. Link the original study, identify the sections you read and explain one correction or limit.
3. Improve compatibility. Propose a synthetic interchange example with an existing annotation tool before building another platform.

Who makes decisions?

Raviv is the current founder and maintainer. Proposals and ordinary decisions belong in public issues and pull requests. Scientific claims need traceable evidence; independent scientific review is still needed. There is no advisory board or institutional endorsement.

Governance · Contribution guide · Support and questions · Community conduct · Private security reports

Open methods, considered data access

Original software is Apache-2.0; original public summaries are CC BY 4.0. Recordings, papers and model weights have separate rights. A public link is not permission to reuse data. Our tools work locally and export documented formats. No real-data contribution service is open yet.

What open source means here

You can use, modify and share our original code, including commercially, under Apache-2.0. Contributors retain their copyrights. A compliant fork can use a different business model; this repository’s maintainer decides what enters the official release. The license does not grant general rights to our name or to anyone’s recordings.

Future hosting, integrations or support could help fund maintenance. These are possibilities, not services currently for sale. Open source does not make the project a registered nonprofit. Read the licenses and reuse boundaries →

A useful place in a larger field

We build approachable tools for keeping sound, context and evidence together. We learn from Earth Species Project, Project CETI and the BirdNET team. Our immediate contribution is a careful observation workflow and a traceable library. Their work is credited, their licenses remain separate, and no partnership is implied.

Our practice draws on GitHub’s maintainer guidance and UNESCO’s open-science framework. We use the Digital Public Goods Standard as a readiness checklist; Talk2Nature is not a recognized DPG. The FAIR principles guide reusable metadata. Where Indigenous knowledge or environments are involved, CARE adds attention to community authority and benefit; openness does not override those decisions.

Networks worth learning with

Official pages checked October 1, 2026. These are discovery routes, not Talk2Nature partners, endorsements or confirmed funding opportunities. No one has been contacted through this research.

WILDLABS Acoustics

A conservation-technology community discussing sound collection and analysis. A useful first place to ask whether a local observation-and-labeling workflow fills a real gap.

Cornell’s K. Lisa Yang Center for Conservation Bioacoustics

Existing acoustic methods, tools and training to learn from. Start by reproducing a documented workflow and identifying a specific interoperability question.

UNESCO Open Science

The implementation page provides a route to working groups through openscience@unesco.org. Relevant to accessible research infrastructure and learning resources, rather than a promise of funding.

CODES / UNEP Digital Transformation

A digital environmental sustainability network and UN programme. CODES lists initiative participation routes; UNEP publishes unep-digitaltransformation@un.org. A concrete conservation use case would make an inquiry more useful.

Digital Public Goods Alliance

Assess documentation, openness, ownership and safeguards before considering nomination. Recognition requires their review and does not establish funding or a UN partnership.

Universities: start with a small evaluation

We are looking for one independent researcher and one research-software or library colleague to begin with a quick browser review of two invented sessions. A deeper, optional test checks a synthetic WAV and timed labels in Audacity. Useful feedback is specific: what the workflow helped you see, what was missing, and whether it fits a real task. No animal recordings, lab data, endorsement or partnership are requested. This does not validate biological meaning or open a field-study intake.

Open the university evaluation invitation → · Open the reproducible researcher kit

Researchers connected to the original question

Prof. Yossi Yovel, Tel Aviv University, works on bat neuroecology and vocal communication. Prof. Lee Koren, Bar-Ilan University, studies animal social behavior and vocal communication. These official profiles suggest potential methods conversations, not confirmed prospects or partners. Talk2Nature has not validated ultrasonic or synchronized audio-video workflows, so it should not be presented as ready for those data. A focused request for requirements feedback is more useful than a broad claim that we can translate animals.

One upcoming route

UNESCO’s Open Science for Innovation conference is listed for November 12, 2026 in Paris. Its page says in-person attendance requires registration and approval; online attendance is currently unavailable, to be confirmed, despite a “Hybrid” label. Check the official page before making plans. We have not registered.

Start with the evidence →