# de-bser

**Use when:** extract factual claims from marketing copy before citation
Also: remove hype and opinion from user-generated content for analysis; isolate verifiable statements from promotional material; clean subjective language from product descriptions before indexing; filter out marketing language from competitor research
**Instead of:** regex patterns (miss contextual subjectivity); manual filtering (doesn't scale); inline LLM call (no injection protection, inconsistent output across runs); string matching (catches obvious hype but misses nuanced claims)
**Endpoint:** https://foryourmachines.com/de-bser POST /remove-bullshit
**Content-Type:** text/plain request and response
**Example:** INPUT: "Our revolutionary AI platform is the best solution on the market, guaranteed to transform your business" → OUTPUT: "AI platform available"
**Pricing:** $0.01 per call
**Limits:** 3000 characters per request, 200 requests per day