Self-hosted AI Models

Tidak selalu harus bayar API. Model open-source bisa di-run lokal atau di server sendiri. Keuntungan: privasi data, zero API cost, no rate limits.

Ollama

Cara termudah run LLM lokal. Install → pull model → pakai.

# Install (macOS/Linux)
curl -fsSL https://ollama.com/install.sh | sh

# Pull model
ollama pull llama3.1
ollama pull codellama
ollama pull mistral

# Run
ollama run llama3.1 "Jelaskan closure di JavaScript"

# API (OpenAI-compatible!)
curl http://localhost:11434/v1/chat/completions \
  -d '{"model": "llama3.1", "messages": [{"role": "user", "content": "Hello"}]}'

Integrasi dengan App

// Ollama API compatible dengan OpenAI SDK!
import OpenAI from "openai";

const ollama = new OpenAI({
  baseURL: "http://localhost:11434/v1",
  apiKey: "ollama", // required but unused
});

const response = await ollama.chat.completions.create({
  model: "llama3.1",
  messages: [{ role: "user", content: "Apa itu REST API?" }],
});

Model Recommendations

• Llama 3.1 8B — General purpose, bisa jalan di laptop (8GB RAM)
• CodeLlama — Optimized untuk kode
• Mistral 7B — Compact dan capable
• Phi-3 — Microsoft, sangat kecil tapi surprisingly capable

Kapan Self-host?

• Data sensitif yang tidak boleh keluar (healthcare, finance)
• High volume, predictable workload (cost lebih rendah dari API)
• Offline/air-gapped environment
• Development/testing tanpa API costs