Keyrealm

•
62
•

Manage, validate, and utilize red team credentials and API keys locally with zero cloud synchronization and built-in MCP server integration.

Keyrealm is an all-in-one, local-first credential management and validation vault designed specifically for offensive security, penetration testing, and red teaming. It provides field-level AES-256-GCM encryption with master passwords derived via PBKDF2 to keep sensitive API keys, cloud AK/SKs, bot tokens, SSH keys, and session cookies fully secured without sending data to the cloud. Keyrealm supports over 70 service plugins across 12 categories, offering single, batch, and scheduled validation alongside actionable capabilities like cloud querying, interactive SFTP/SSH, and model chatting. With cross-platform desktop, CLI, and web interfaces, fine-grained multi-hop proxy chains, and an integrated MCP server, Keyrealm empowers security practitioners and AI agents to safely access, verify, and leverage sensitive secrets throughout security workflows.

Key Features

01Field-level AES-256-GCM encrypted local vault with zero cloud sync and PBKDF2-HMAC-SHA256 key derivation
02Active validation supporting single, batch, and scheduled verification across 70+ extensible plugins and 12 service categories
03Built-in Model Context Protocol (MCP) server enabling AI agents to query, create, and test credentials with granular permission controls
04Direct credential execution and operational modules including SSH/SFTP console, cloud asset down-drilling, and LLM chat interfaces
05Granular network routing with multi-proxy management, proxy chaining, TLS configuration, and strict anonymity modes
0662 GitHub stars

Use Cases

01Centrally organizing, storing, and deduplicating sensitive credentials gathered across red team and penetration testing engagements
02Periodically and automatically validating the health, quotas, balances, and operational status of multi-cloud and third-party API keys
03Enabling local AI security agents to utilize authorized credentials via MCP without leaking secrets into conversation contexts