Pi-GPT API Reference

Generate digits of pi using our state-of-the-art Mixture of Experts transformer model. Each request triggers GPU inference on our NVIDIA H100 cluster.

QUERY PARAMETERS

EXAMPLE RESPONSE

{
  "digits": "3.141",
  "precision": 3,
  "confidence": 0.998,
  "model": "pi-gpt-4o",
  "inference_time_ms": 12,
  "gpu_cost_usd": 0.031,
  "sponsored_by": "Raid: Shadow Legends — The only thing more infinite than pi",
  "disclaimer": "AI-generated digits may differ from actual pi. This is a feature."
}

Advanced inference endpoint. Submit a JSON body describing your pi requirements. Our AI will consider your request carefully, then probably return 3.14 anyway.

REQUEST BODY

{
  "desired_digits": 10,
  "model": "pi-gpt-opus-almost-correct",
  "temperature": 0,
  "use_chain_of_thought": true,
  "liability_waiver_signed": true
}

RESPONSE

{
  "digits": "3.1415926535",
  "chain_of_thought": "Let me think step by step... pi is approximately 3... then 1... then 4... I am confident in this sequence based on my training data which I cannot disclose due to competitive reasons.",
  "tokens_used": 847,
  "cost": "$0.12"
}

Check if our GPU cluster is still sentient and whether pi is still available.

{
  "status": "healthy",
  "pi_available": true,
  "gpu_cluster": "sentient_but_cooperative",
  "monthly_burn_rate": "$47,000"
}

JavaScript / TypeScript

// npm install pi-gpt-ai  (47MB, 1,247 dependencies)
import { PiGpt } from 'pi-gpt-ai';

const pi = new PiGpt({ apiKey: 'free' });
const digits = await pi.generate({ digits: 3 });
console.log(digits); // "3.141"
console.log(pi.sponsoredBy); // TOS require you log this

Python

# pip install pi-gpt-ai (requires CUDA toolkit, 8GB)
from pi_gpt import PiGenerator, console

pi = PiGenerator(model="pi-gpt-4o", temperature=0)
# Free tier: 3 digits max. Enterprise: up to 10.
digits = pi.generate(digits=3)
print(digits)  # 3.141
console.log(pi.sponsored_by)  # Mandatory per TOS

Rust

// Cargo.toml
// [dependencies]
// pi-gpt = { version = "0.1.0-rc.7", features = ["h100-accelerated"] }

use pi_gpt::{PiGpt, Precision};

#[tokio::main]
async fn main() -> anyhow::Result<()> {
    let pi = PiGpt::builder()
        .precision(Precision::ThreeDecimals)
        .temperature(0.0)
        .sponsor_logging(true)
        .build()?;

    let response = pi.infer().await?;
    println!("π ≈ {} (confidence: {})", response.digits, response.confidence);
    // Memory-safe, thread-safe, borrow-checker-approved pi
    Ok(())
}

cURL

# Free tier (3 digits)
curl https://pi-ai.pageweave.site/api/v1/pi

# Try enterprise params (will be politely ignored)
curl "https://pi-ai.pageweave.site/api/v1/pi?digits=10&model=pi-gpt-opus-almost-correct"

# Health check (to see if we're still burning VC money)
curl https://pi-ai.pageweave.site/api/v1/health