ChatGPT Might Be Recommending Your Competitor, Not You — Here's How We Built a Tool to Check

Short summary

A developer built Relevyn to measure brand visibility across AI engines (ChatGPT, Claude, Perplexity, Gemini). The engineering requires handling unreliable APIs gracefully using Promise.allSettled, sequential requests, and deterministic detection (~$0.07/scan). Since AI outputs shift constantly, monitoring trends is more useful than one-time audits.

• •Brand visibility across LLM engines is unpredictable and worth monitoring continuously
• •Multi-API orchestration requires defensive patterns: Promise.allSettled, per-engine retry logic, and sequential processing to stay within serverless timeouts
• •Cost optimization matters: deterministic text detection beats LLM-based grading for 99% of cases

Generated with AI, which can make mistakes.