I Built a Programmatic SEO Site with Astro & DeepSeek API (300+ Pages for $1.65

#astro #python #webdev #showdev

The Problem: AI Fatigue 🤯

There are too many AI tools. I wanted to build a "Referee" site that compares them (e.g., ChatGPT vs Claude, Notion vs Obsidian) and gives a definitive verdict.

But I had two constraints:

Zero Budget: I didn't want to pay for a database or expensive hosting.

Scalability: I needed to generate 300+ pages, but I didn't want to write them manually.

The Solution: The "pSEO" Stack 🛠️

I chose a JAMstack architecture focused on speed and low cost:

Frontend: Astro (SSG mode)

Styling: Tailwind CSS

Content Engine: Python + DeepSeek V3 API

Database: A single JSON file

Hosting: Vercel (Free Tier)

Here is the live demo: PrismKit.xyz

Step 1: The "Database" (JSON) 📂

Instead of setting up Postgres or MongoDB, I just used a flat JSON file. It's fast, free, and version-controllable.

// tools_database.json
[
{
"id": "notion",
"name": "Notion",
"category": "Productivity",
"alternatives": ["obsidian", "evernote"]
},
{
"id": "obsidian",
"name": "Obsidian",
"category": "Productivity",
"alternatives": ["notion"]
}
]

Step 2: The Content Factory (Python + DeepSeek) 🤖

This is the magic part. I needed unique comparisons for every pair. Writing them by hand is impossible. Using GPT-4o is expensive.

I used DeepSeek V3 via their API (OpenAI compatible). It's incredibly cheap (~$0.05 for the whole site).

Here is the simplified logic of my content_factory.py:

from openai import OpenAI
import json
client = OpenAI(
api_key="sk-...",
base_url="https://api.deepseek.com"
)
def generate_verdict(tool_a, tool_b):
prompt = f"Compare {tool_a} and {tool_b}. Who should use which? Be concise."
response = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": prompt}]
)

return response.choices[0].message.content
Loop through tools and save to JSON...

The script runs locally, updates the JSON file with the AI-generated insights, and then I push the JSON to GitHub.

Step 3: Dynamic Routing in Astro 🚀

Astro makes Programmatic SEO (pSEO) trivial with getStaticPaths. I created one file: src/pages/vs/[...slug].astro.

// src/pages/vs/[...slug].astro
export async function getStaticPaths() {
const tools = await import('../../data/tools_database.json');

// Generate all possible pairs
const paths = [];
tools.forEach(tool => {
tool.alternatives.forEach(altId => {
paths.push({
params: { slug: ${tool.id}-vs-${altId} },
props: { toolA: tool, toolB: getToolById(altId) }
});
});
});
return paths;
}
const { toolA, toolB } = Astro.props;