Self Chat Phone

A single-page phone-style chat that lets two virtual contacts carry on an autonomous text conversation using any local Ollama model. Pick a model from the dropdown and watch Alex and Blair volley short, natural replies forever.

Prerequisites

• Ollama running locally (ollama serve).
• CORS access from the browser to Ollama. If your browser blocks requests, launch Ollama with OLLAMA_ORIGINS=* (for example: OLLAMA_ORIGINS=* ollama serve).

Run It

1. Open this project in Pinokio.
2. Launch the app (Pinokio will serve index.html).
3. Select a model from the dropdown and let the chat flow. Switching models resets the conversation with the new model instantly.

The interface mimics a modern phone SMS app, complete with a live status indicator and an iPhone-inspired tri-tone chime (tap/click once to unlock audio). Alex and Blair keep their replies to short, readable bursts so you can follow along in real time.

API Cheatsheet

Use these endpoints on your local Ollama instance. When the app is served over HTTPS, Ollama must be accessed at https://11434.localhost; otherwise use http://localhost:11434.

JavaScript

const base = window.location.protocol === "https:"
  ? "https://11434.localhost"
  : "http://localhost:11434";
const models = await fetch(`${base}/api/tags`).then((r) => r.json());

const reply = await fetch(`${base}/api/generate`, {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "llama3",
    prompt: "Write a single friendly chat reply.",
    stream: false
  })
}).then((r) => r.json());

Python

import requests

# Swap to "https://11434.localhost" if your client must use HTTPS
base = "http://localhost:11434"
models = requests.get(f"{base}/api/tags").json()

payload = {
    "model": "llama3",
    "prompt": "Write a single friendly chat reply.",
    "stream": False,
}
reply = requests.post(f"{base}/api/generate", json=payload).json()

curl

# Swap to https://11434.localhost when calling from an HTTPS-only context
curl http://localhost:11434/api/generate \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "llama3",
    "prompt": "Write a single friendly chat reply.",
    "stream": false
  }'

Notes

• Refresh the page if you want to clear the conversation manually.
• If you see connection errors, confirm Ollama is running and accessible at http://localhost:11434 (or https://11434.localhost when browsed over HTTPS) with permissive CORS settings.
• If Ollama is offline you'll see an in-app red banner reminding you to start ollama serve.