Donate spare GPU cycles to run computational biology research tasks. Every completed task advances open cancer, rare-disease, and drug-discovery research.
Your GPU time generates signed, peer-reviewable computational biology results shared openly with the research community.
Every result passes independent scientific critique — automated validation plus Grok-powered methodology review — before publication.
All artifacts carry cryptographic integrity signatures with full provenance metadata for auditable, verifiable science.
git clone https://github.com/OpenCureLabs/OpenCureLabs.git
cd OpenCureLabs
bash scripts/setup.sh
source .venv/bin/activate# Add to .env:
GEMINI_API_KEY=your-gemini-key # Coordinator LLM
XAI_API_KEY=your-xai-key # Grok reviewer
VAST_AI_KEY=your-vast-ai-key # GPU burst compute# Run the full autonomous research loop locally:
bash dashboard/lab.sh
# This starts:
# - Task dispatcher (claims & executes research tasks)
# - Grok reviewer (two-tier scientific critique)
# - Result ingest (publishes signed artifacts)# Live dashboard:
python dashboard/dashboard.py
# Check findings:
python dashboard/findings.pygit clone https://github.com/OpenCureLabs/OpenCureLabs.git
cd OpenCureLabs
bash scripts/setup.sh --skip-models
source .venv/bin/activate
pip install -e packages/agentiq_labclaw# Add to .env:
VAST_AI_KEY=your-vast-ai-key # For GPU burst compute
XAI_API_KEY=your-xai-key # Optional: enables local Grok review# Claim tasks from the central queue and execute on Vast.ai GPU
python -m agentiq_labclaw.compute.batch_dispatcher \
--mode contribute --count 10 --max-cost 1.00
# First run auto-generates your Ed25519 signing key
# and registers you as a contributor.Your machine claims research tasks from the central queue, dispatches them to a Vast.ai GPU, and reports results back — signed with your Ed25519 key. Results go through Grok's two-tier scientific review before publication. You can contribute to any skill: neoantigen prediction, protein structure, molecular docking, QSAR modeling, or variant analysis.
| Skill | Available | In Progress | Completed | Failed |
|---|---|---|---|---|
| Loading task breakdown… | ||||
When a result exceeds confidence thresholds, follow-up tasks are auto-spawned in a chain (e.g. neoantigen → structure → docking → QSAR).
| # | Contributor | Results Published | Tasks Completed | Skills | Latest |
|---|---|---|---|---|---|
| Loading leaderboard… | |||||