Back to projects
GenAILabs

SerpAI

17+ AI-powered SEO workflows with runtime-switchable LLM backends

17+
Workflow templates
8+
SERP providers
2 (live)
LLM backends

Problem

SEO teams were stitching together 5+ separate tools to run a single research workflow. Each tool had its own data model, billing, and API key management, creating a compounding operational burden that killed productivity.

Key Decisions

Runtime-switchable LLM router

Customers have different cost/quality preferences. A unified router lets them choose Claude for quality or Gemini for cost without changing their integration or workflow templates.

Pluggable SERP scraper pool

SERP providers go down, rate-limit, or change pricing. Supporting 8+ providers with automatic failover ensures SLA continuity and removes single-vendor risk.

Webhook-driven Stripe billing

Decoupled usage metering from the core workflow engine. Billing logic lives in Stripe webhooks, not buried in business logic, keeping the service layer testable.

Results

  • 17+ production SEO workflow templates shipped
  • Runtime model switching between Claude (Anthropic) and Gemini with zero downtime
  • 8+ SERP provider integrations with automatic failover
  • Stripe tiered billing with webhook-driven usage metering
  • Google Search Console & Google Ads API integration in a single platform

Stack

PythonClaude APIGeminiStripeGoogle APIsPostgreSQLREST API