Skip to main content
Get answers to common questions about designing, launching, and operating Crunches on the Crunch Protocol.
Coordinators design, launch, and operate “Crunches” — decentralized prediction challenges that tap into the talent of 10,000+ ML engineers and 1,200+ PhDs. Coordinators bring data, define the problem, and set incentives. Crunch Labs provides the infrastructure, orchestration, and global contributor network.
Yes! Coordinators can run:Continuous (ongoing) Crunches
  • New rounds daily/weekly
  • Continuous predictions streamed in real-time
  • Ideal for price data, risk, or volatility forecasting
Time-boxed Crunches
  • Fixed start/end
  • Ideal for research, Kaggle-style problems, or academic studies
Continuous Crunches are supported natively by the Crunch Engine and real-time modules.
Coordinators retain access to:
  • All model outputs generated during the Crunch
  • Rankings, metrics, and aggregated performance reports
  • Residual error analyses and feature importance (if enabled)
  • Coordinator dashboards for post-competition analysis
If you want ongoing predictions after the initial Crunch ends, you can:
  • Switch to continuous mode
  • Launch a “Season 2”
  • Or integrate predictions via Crunch’s API
This is common for institutions that want to transition from research → production.
It depends on the Crunch type:A. Using Public Market Data (e.g., crypto, equities, FX)Crunch Labs can support you with:
  • Real-time price feeds (if using supported sources)
  • Secure ingestion pipelines
  • Normalized data formats ready for modeling
If you want custom datasets, the Coordinator must provide:
  • Historical datasets
  • Clearly documented features/targets
  • Evaluation rules
Crunch Labs helps with formatting, preprocessing, and secure delivery to Crunchers.B. Using Proprietary Enterprise Data Coordinators may provide:
  • Fully anonymized datasets
  • Differential Privacy–wrapped features
  • Synthetic mirrors of sensitive data
Crunch supports:
  • Secure compute (MPC, TEE)
  • Encrypted submissions
  • Private Crunches (invite-only or closed-scope)
Base Model
  • Coordinators provide the prize pool, paid out in USDC.
  • Crunch Labs handles distribution, scoring, and compliance.
Co-Funding / Grants (Coming Soon) Yes — Crunch Labs can co-fund or grant-match high-value Coordinators, especially for:
  • Real-time prediction markets (crypto, macro, FX)
  • Institutional research
  • Healthcare, risk, or environmental forecasting
  • High-impact ML/AI challenges
  • Early CIP Seasons
Grant matching is decided based on impact, dataset uniqueness, and strategic relevance.
Crunch Labs provides:
  • The Crunch Engine (orchestration layer for ML competitions)
  • Model submission & scoring pipelines
  • Privacy tools: Differential privacy, MPC, and secure enclaves
  • Dataset normalization & distribution
  • Contributor network (10k+)
  • Real-time evaluation modules
  • Coordinator dashboards
  • USDC prize distribution
  • Support for Solana + Ethereum environments
Coordinators only need to focus on defining the problem and providing datasets or feeds.
  1. Submit the Coordinator Application
  2. Meet with Crunch Labs to refine the challenge
  3. Provide dataset or choose built-in feeds
  4. Approve scoring metrics & schedule
  5. Set prize pool and duration
  6. Launch the Crunch to the global network
The Crunch Coordinator Program offers additional support, emissions, and increased visibility.