> For the complete documentation index, see [llms.txt](https://openvision-2.gitbook.io/whitepaper/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://openvision-2.gitbook.io/whitepaper/modular-compute-protocol.md).

# Modular Compute Protocol

<figure><img src="/files/RPQHWkzrKEuKQ6WN4SB2" alt=""><figcaption></figcaption></figure>

MCP powers decentralized AI training at scale:

* **Task Assignment:** Micro-tasks are dynamically matched to VisionNodes
* **Reputation System:** Contributor scores based on task success, uptime, latency
* **Task Verification:** Quorum-based or zkML-based proof of result integrity. If quorum
  \
  fails, zk proof is required for validation.
* **Resource Optimization:** Scheduler avoids node overload and ensures fairness
* **Slashing Logic:** Faulty or malicious contributors are penalized via token slashing
  \
  and reputation decay