RemoteX is a local-first AI workspace designed for remote infrastructure operations, enabling users to investigate and fix production issues directly from their phone or desktop. It seamlessly connects to servers, clusters, databases, logs, files, and Docker applications, presenting evidence and proposing fixes that require explicit user approval before execution. This ensures a safe and controlled approach to issue resolution.

• Investigate and Fix Issues: Safely chat with your servers to diagnose problems, get proposed fixes, and verify the results.

• Broad Connectivity: Supports direct connections to SSH servers, Docker hosts, Kubernetes clusters, and various databases (Postgres, MySQL, SQLite, Redis) without requiring agents or proxies.

• Bring Your Own AI: Integrates with 14 AI providers including OpenAI, Gemini, Anthropic, and local models via Ollama and LM Studio, allowing users to leverage their existing subscriptions or keys. RemoteX never proxies AI calls, ensuring privacy.

• Unified Workspace: Consolidates investigations into a single thread, attaching servers, clusters, databases, logs, and files for a comprehensive view. Projects group related infrastructure for better organization.

• Activity and Results: View command output directly in the terminal flow and browse database schemas and data within the chat interface.

• Snippets and Playbooks: Save frequently used commands or create multi-step routines with approval gates for efficient workflows.

• Plan Mode: Allows the AI to propose an investigation plan first, which the user can then approve before execution.

• Mobile and Desktop Sync: End-to-end encrypted synchronization across Mac, iPhone, and iPad via iCloud Drive, with a passphrase only the user knows.

• Direct Connections & Security: Credentials (SSH keys, AI keys, etc.) are stored securely on the device using OS secure stores. All connections, including AI prompts, go directly from the user's device to the target infrastructure or AI provider.

• Approvals and Verification: Every state-changing action requires explicit biometric or local approval. Read-only checks run autonomously, and verification steps confirm the success of applied fixes.