**ClaudeAPI** wraps the Claude API with enterprise-grade security controls—key scrubbing, request fingerprinting, anomaly detection, and Docker isolation—protecting sensitive data across logs, outputs, and network traffic. Built with Python, cryptography, and Docker with non-root containers and network isolation to Anthropic endpoints only. Includes a full security audit suite, usage dashboard, rate limiting (60 RPM default), and device fingerprinting that catches unauthorized access patterns.
RizzBrain aggregates ChatGPT dating conversations into a searchable, voice-learned archive that summarizes relationship status, drafts contextual replies, and surfaces pattern insights ("where do chats die?"). Built on Flask + Claude (Haiku/Sonnet for speed/quality tradeoff), it ingests via iOS Shortcut share-sheet and serves a mobile-first dark UI, deployed live at `rizz.thirstyai.live` handling real user data with persona-based coaching.
**ThirstyPrompt** turns vague project ideas into production-ready enterprise prompts via an interactive Telegram bot that asks targeted questions about architecture, stack, and deployment, then synthesizes responses into comprehensive AI prompts using Claude Haiku. Built with python-telegram-bot and Anthropic's API, it handles Telegram's 4096-char limit through chunked message delivery and persists completed builds to SQLite. Generates up to 8,000-token master prompts covering architecture decisions, data models, testing strategy, and risk analysis in a single conversation flow.
NEED ONE BUILT FOR YOU?
San Antonio bar operator who codes. Days, not weeks.
✓ 5 hours per week back, or 100% refund