COCOON enables running AI models in trusted execution environments, while earning TON cryptocurrency for compute services.

• GPU owners earn TON by serving models
• App developers plug into low-cost, secure and verifiable AI compute
• Users enjoy AI seamlessly, with full privacy and confidentiality

This repository contains all the necessary tools and documentation to both serve and access models via COCOON.

For Workers: offering GPUs for computation

• Download: Latest worker release – Ready-to-run TDX image and setup scripts
• Setup Guide: Instructions are included in the release archive (preview here)

For Developers: requiring secure AI compute

• Build Instructions: See below for reproducing worker distribution from source

Anyone can verify the worker distribution by rebuilding from source. Note that this step is not needed to run your own workers.

# 1. Build the VM image (reproducible)
./scripts/build-image prod

# 2. Generate distribution
./scripts/prepare-worker-dist ../cocoon-worker-dist

# 3. Verify the TDX image matches the published release
cd ../cocoon-worker-dist
sha256sum images/prod/{OVMF.fd,image.vmlinuz,image.initrd,image.cmdline}
# Compare with the published checksums

The same goes for model images:

# 1. This will generate a model tar file with the full model name, which includes hash and commit.
./scripts/build-model Qwen/Qwen3-0.6B
# Compare with the published model name

See LICENSE file.