You don’t need to know everything.
Just where to start.

From an idea to “hello”.

A chatbot reads a message and sends a reply. You decide its purpose; your Python code connects the pieces.

1. 1

   Choose something you’d like to build

   Open a guided project. Start with Study Buddy, Quiz Master, Story Studio, or Idea Lab.

2. 2

   Give it a name and a personality

   Describe how your bot should help. The guided builder turns your choices into four editable Python files.

3. 3

   Run it. Say hello. Make a change.

   Open your project in the playground and select Run my bot. When chat says Ready, send a message. Stop the bot before running changed code again.

Give your bot a way to think.

Select Connect AI in the playground. Choose a provider, use its “Get a key” link, then paste your API key into the labelled field and save.

Saving a key doesn’t verify it. Run your bot and send a message to test the connection. Keys are sent with run requests to Dhiora’s server for provider calls, and provider usage may cost money.

The chat is your testing ground.

Try the suggested questions or write your own. Enter sends a message; Shift + Enter adds a new line. On a phone, switch between Your code and Test your bot.

Select Style in the chat panel to customise colours, backgrounds, and message bubbles. Use the expand control when you want more room.

Four files. A place for every idea.

Start with the personality file. You can explore the other pieces as you get comfortable.

# These instructions shape every answer. Experiment with them!
SYSTEM_PROMPT = "You are a friendly study buddy for students. Explain ideas in simple language, use a relatable example, and ask one question to check understanding. Give hints before giving answers. If you are unsure, say so."

# Give your chatbot a name.
BOT_NAME = "Study Buddy"

# Keep the last three exchanges so your bot can follow a conversation.
# Restarting the bot clears this memory.
from collections import deque
from ai import ask
from prompt import SYSTEM_PROMPT

conversation = deque(maxlen=6)

def respond(message: str) -> str:
    history = "\n".join(conversation)
    answer = ask(f"{SYSTEM_PROMPT}\n\nRecent conversation:\n{history}\nStudent: {message}\nAssistant:")
    conversation.append(f"Student: {message[:2000]}")
    conversation.append(f"Assistant: {answer[:2000]}")
    return answer

# This is the starting point of your chatbot — the file that runs first.
# The real work happens in the other 3 tabs (config.py, prompt.py, logic.py);
# this file just plugs them together. You usually won't need to change much
# here — try editing the other files first!
from ai import *
from config import BOT_NAME
from logic import respond

bot = ChatBot(BOT_NAME)

# Whenever someone sends your chatbot a message, Python runs this function
# and sends whatever it returns back as the reply.
@bot.on_message
def handle_message(message):
    return respond(message)

# This line switches your chatbot on. Keep it at the very bottom of the file.
bot.start()

A little code. A lot you can do.

These predefined functions save you from writing everything yourself. Expand a function for its explanation and examples.

11 functions found

bot = ChatBot()

@bot.on_message
def respond(message):
    return "Hello " + message

bot.start()

answer = ask("What is the capital of France?")

@bot.on_message
def respond(message):
    return reply(message)

mem = memory()
mem.append("user", "hi")

db = database()
db.set("visits", db.get("visits", 0) + 1)

save("high_score", 42)

score = load("high_score", 0)

users = create_table("users")
users.insert({"name": "Ada"})

query("users", name="Ada")

Your next “what if?”

Stuck? You’re still learning.

Small words, big ideas.

Prompt

Instructions that tell an AI what you want it to do.

Function

A reusable set of steps, such as ask().

Variable

A name for a value, like BOT_NAME holding your bot’s name.

String

Text wrapped in quotation marks.

API key

A private credential that lets your code use an AI provider.

Run

Execute your code to see what it does.