How I Built and Launched 2 Commercial APIs in a Single Day

#api #python #fastapi #sideprojects

I've been seeing developers make money selling APIs on marketplaces like RapidAPI, and I wanted to try it myself. Here's exactly how I went from zero to two live API products in one day.

The APIs

TextPulse — A text analytics API with 7 endpoints:

Sentiment analysis with per-sentence breakdown
Keyword extraction with relevance scoring
Readability scoring (Flesch Reading Ease + Flesch-Kincaid)
Named entity extraction (emails, URLs, phones, money, dates)
Language detection
Extractive text summarization
Full analysis (all of the above in one call)

ShieldCheck — A password & security API with 11 endpoints:

Password strength analysis (0-100 scoring with pattern detection)
Crack time estimation across 5 attack scenarios
Secure password generation
Password hashing (SHA-256/512)
Email validation + disposable domain detection
Domain breach checking
TOTP 2FA generation and verification
Secure token generation
Shannon entropy calculation
Full security audit

The Stack

Both APIs use the same simple stack:

Python + FastAPI for the API
Pydantic for request/response validation
Railway (TextPulse) and Render (ShieldCheck) for hosting
RapidAPI as the marketplace
GitHub for version control

No machine learning libraries. No external API dependencies. No databases (yet). Everything runs as pure Python logic, which keeps deployment simple and responses fast.

The Build Process

The actual coding was the fastest part. Each API is a single server.py file with:

Input models (Pydantic)
Core analysis functions
API key validation
FastAPI endpoints

The time-consuming part was everything else:

Configuring RapidAPI (adding each endpoint, setting media types, writing example bodies)
Setting up pricing tiers
Writing documentation
Testing every endpoint through the marketplace

Pricing Strategy

I researched competitors on RapidAPI before setting prices. Key insight: most text analytics APIs only do ONE thing (just sentiment, or just keywords). TextPulse bundles 7 features, so I can price below the sum of individual tools while still being competitive.

Both APIs have 4 tiers: Free (tight limits for testing), Pro (recommended, best value), Ultra, and Mega.

What I'd Do Differently

Start with documentation — I wrote it after building, but good docs should be planned alongside the API design
Test through RapidAPI earlier — There were media type configuration issues that only showed up when testing through the marketplace
Pick Python 3.12 from the start — Python 3.14 had dependency compatibility issues that cost me time