Skip to main content

What It Is

One Binary, Complete Stack What others assemble from six services, this ships in 50MB of Rust. PostgreSQL is the only dependency.

How You Run It

Deploy Anywhere The same binary runs in Docker, on bare metal, in the cloud, or in an air-gapped network. One CLI, Every Environment Manage agents, servers, and deployments from your terminal across local, staging, and production.

Why You Can Trust It

Every Layer, An Open Standard MCP, OAuth 2.0, WebAuthn, Git, ChaCha20-Poly1305. Zero proprietary protocols at any layer.

Own your Cowork deployment

Claude Cowork on Your Infrastructure Your /v1/messages gateway, signed MCP allowlist, and org-plugins supply chain. Route inference to any upstream you operate. Every tool call governed, every prompt logged, nothing leaves your network. CTO / CISO Brief The strategic case for Cowork-on-your-infrastructure. Why Anthropic's April 2026 opening is a procurement inflection, the five things you now own, RFI/RFP question bank, and a 90-day delivery plan. Enterprise Factsheet (PDF) Two-page brief: govern and own Claude Cowork on self-hosted infrastructure. For CISO/CTO review.

Technical documentation

Deployment Reference Five-move architecture, three auth tiers (PAT, session, mTLS), Ed25519 signed manifest, audit schema. The platform-agnostic reference that underpins the install guides. Install on macOS .mobileconfig payload, Jamf and Intune for Mac, Secure Enclave mTLS. Step-by-step. Install on Windows Registry policy, Intune Custom OMA-URI, Group Policy ADMX, TPM-backed mTLS. Step-by-step. Gateway Service /v1/messages and /v1/auth/cowork/* endpoint reference, provider routing, MCP allowlist routes.

What It Does

Governance Pipeline Four enforcement layers evaluate every tool call before execution. Scope, secrets, rate limits, and audit. Secrets Management Credentials are encrypted at rest and blocked from inference. Secrets never reach AI models. MCP Governance Every MCP server is registered, authenticated, and governed. Per-server OAuth2 with full audit trail.

How It Protects You

Analytics & Observability Every AI request is logged with tokens, cost, latency, model, user, and agent. Exportable to your SIEM. Closed-Loop Agents Agents monitor their own performance, detect drift, and adapt automatically.

Why It Passes Audit

Compliance SOC 2, ISO 27001, HIPAA, and OWASP Agentic Top 10. Built to survive an audit.

What It Connects To

Any AI Agent Claude, Codex, Gemini, or your own custom agents. Provider-agnostic governance for all of them.

How You Use It

Claude Desktop & Cowork One-click governed skills inside Claude. The fastest path from install to governed AI. Web Server & Publisher The same binary serves your website, blog, and documentation. CMS and static site generation built in.

Why It Scales

Fully Extensible Architecture Add routes, tables, jobs, and providers through Rust extension traits. Your code compiles into your binary.

Learn

Documentation Deployment guides, API reference, and configuration for every domain. Guides Step-by-step tutorials, workflows, and best practices for AI governance. Live Demo Watch four governance layers execute in real time in under six minutes. About Founder-led startup building self-hosted AI governance infrastructure.

Login

Blog

Articles and updates

Sort:

Example Blog Post

This is an example blog post demonstrating the content system. Replace it with your own content.

January 01, 2025

Infrastructure

Self-Hosted Deployment Deploy Anywhere Unified Control Plane No Vendor Lock-In

Capabilities

Governance Pipeline Secrets Management MCP Governance Analytics & Observability Closed-Loop Agents Compliance

Integrations

Any AI Agent Claude Desktop & Cowork Web Server & Publisher Extensible Architecture

Resources

Documentation Guides Enterprise Factsheet (PDF) Sitemap llms.txt Login

Legal

Privacy Policy Cookie Policy Terms of Service Refund Policy

Contact

Discord LinkedIn GitHub

© 2026 systemprompt.io