OpenClaude FAQs

Question 1

What is OpenClaude?

Accepted Answer

OpenClaude is a local industrial control plane designed to execute the leaked 'Claude Code' agentic harness using open-weight AI models via Ollama. It promotes technical sovereignty by eliminating proprietary cloud inference dependencies.

Question 2

What are OpenClaude's core features?

Accepted Answer

Key features include a standardized FastMCP 3.2 control plane, zeropaid inference optimized for local execution with Ollama, real-time model hot-switching, the KAIROS system for background memory consolidation, and a React-based industrial oversight dashboard.

Question 3

What is the 'Claude Code leak' mentioned by OpenClaude?

Accepted Answer

The '2026 Claude Code leak' refers to the exposure of an agentic harness that orchestrated AI tool use. OpenClaude leverages this leak to decouple agentic capabilities from specific proprietary inference engines, enabling independent and sovereign AI-assisted engineering.

Question 4

How does OpenClaude eliminate cloud inference costs?

Accepted Answer

It enables 'zeropaid' local inference by routing the Claude Code's agentic logic to open-weight models (e.g., Gemma, Qwen) managed through Ollama, thereby removing the need for recurring API overhead from proprietary cloud providers.

Question 5

Which AI models are supported for local execution?

Accepted Answer

OpenClaude is optimized for leading open-weight models available via Ollama, specifically recommending Gemma 4 26B-A4B for primary execution and Qwen 3.5 (MoE and Dense versions) for KAIROS operations and complex planning.

OpenClaude

OpenClaude

主要功能

使用案例

主要功能

使用案例