
Neumann is built entirely on a Rust-based technology stack, allowing it to operate with performance that is very close to native execution. Neumann achieves high efficiency without compromising reliability or scalability.
[01]
Neumann understands your entire project, not just the file you're looking at. Agents reason across repositories, dependencies, configs, and tools to make changes that are globally consistent and production-ready.
[02]
Neumann speaks the Agent Client Protocol, allowing it to plug into best-in-class agents like Claude Code, Gemini CLI, OpenCode, and future models. No lock-in, no single-model dependency—just the best agent for the job.
[03]
Built on a Rust-first stack, Neumann runs with near-native performance and predictable resource usage. This enables long-running agents, background reasoning, and parallel task execution without slowing down your workflow.
Neumann prioritize protecting users' privacy and data security, adhering to the principle of "local-first" and "minimal data collection".
Codebase files are stored locally. For indexing, files may be temporarily uploaded for embedding. After processing, all plaintext is deleted. You can also use tools like Privacy Mode and the "ignore" function to limit how your data is used.
Strict access control and encrypted transmission are enforced to prevent unauthorized access and reduce exposure to security risks.
User data and infrastructure are deployed based on account location, stored in the United States, Singapore, and Malaysia, with isolation in place to meet local data regulations.