Project ProtoStar

Coalition initiative, with Hugging Face as infrastructure partner, to release publicly funded science data in AI-ready formats.

Hugging Face · Data, training, and fine-tuning platforms (datasets, labeling, RAG/vector DBs, fine-tuning-as-a-service) · open-source

Published in Feed

Overview

Project ProtoStar, announced on October 29, 2026, is a coalition of research institutions, industry and the open science community that helps researchers release datasets, models and code from publicly funded work in machine-actionable, AI-ready formats on open infrastructure. Founding partners include Carnegie Mellon, Purdue, Georgia Tech, Arizona State and George Washington University (via the Open Forum for AI), the Astera Institute, the Open Research Community Accelerator and Hugging Face as infrastructure partner. It provides publishing guidance and a series of open challenges that demonstrate reuse of released data. Its first challenge, hosted by Carnegie Mellon from November 2 to 13, 2026, asks teams to forecast urban climate indicators for 210 U.S. cities.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • Guidance for publishing scientific artifacts in AI-ready, machine-actionable formats
  • Open challenges that show reuse of released data
  • Hugging Face hosts over 17,000 public datasets in core scientific domains (about 29,000 including medicine and mathematics)
  • First challenge: CMU Climate Risk Intelligence Challenge, November 2-13, 2026, training on data through 2018 and predicting 2019-2026 for 210 U.S. cities
  • Partners join through a memorandum of understanding

Use cases

  • Research groups releasing federally funded datasets so ML researchers can reuse them
  • Universities running public forecasting challenges on open scientific data

Pricing

Free to join; contact protostar@orcaopen.org. Challenge results are expected in early 2027.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
29 Oct 2026
Record updated
3 Nov 2026
Record last checked
3 Nov 2026

Explore the Feed

Other entries filed under data, training, and fine-tuning platforms (datasets, labeling, rag/vector dbs, fine-tuning-as-a-service); inclusion does not imply equivalent features.

  • DigitalOcean Agent Data Services

    DigitalOcean's managed data layer for agents: ingestion flows, a catalog with vectors, and SQL or vector query, in private preview.

  • pplx-embed-v2-late

    Perplexity's late-interaction multimodal embedding models (0.6B and 9B) with 128-dim per-token vectors, open on Hugging Face.

  • EmbeddingGemma 2

    Open-weight 740M-parameter multimodal embedding model built on Gemma 4 that encodes text, images, audio, and video for on-device search.

Browse the Feed