Spark is your AI thinking coach. It won't give you the answers — it will help you find them yourself. Here's everything you need to get started.
Log in to SparkFirst things first
Every time you send a message, Spark will respond with a question, a hint, or a challenge — not a ready-made answer. That might feel frustrating at first, but it is how real learning works.
Your teacher can read all your conversations with Spark. This is so they can help keep you safe and see how you are getting on. Spark also has safety filters — if something inappropriate comes up, it will not respond and your teacher will be notified.
Session Modes
When you start a new chat, pick the mode that matches what you are working on. Each one coaches differently.
Scratch
Block-based coding
Use this mode when you're building a Scratch project. Spark won't tell you which blocks to click — instead it'll help you work out the logic yourself. Expect questions like "What do you want your sprite to do when the green flag is clicked?"
Tip: Describe what's going wrong in plain English. The more specific you are, the better Spark can help.
Python
Python programming
Use this mode for Python coding. Spark will ask you to explain what you want your code to do before writing anything — then help you build it one step at a time. You'll always be asked to explain each line before moving on.
Tip: Don't paste your code and ask Spark to fix it. Instead, describe the problem — you'll learn far more.
Micro:bit
Micro:bit & MicroPython
Use this mode when coding a BBC micro:bit. Spark coaches you through MicroPython one step at a time, and always thinks about the physical hardware — what will the device actually do when this runs?
Tip: Flash your code and tell Spark what actually happened. Real results are better than guessing.
3D Design
Tinkercad or OpenSCAD
Use this mode for 3D design work. Spark will first ask whether you're using Tinkercad (drag and drop) or OpenSCAD (code-based). It coaches you on the concepts — not a click-by-click walkthrough.
Tip: Describe your design in plain English first. What shape is it? What is it for? Start there.
Prompt Practice
Writing better prompts
Use this mode to practise writing AI prompts. You'll submit a draft, Spark will analyse it against a framework (Role, Context, Task, Format, Constraints), ask you questions, and help you improve it — before it ever gets sent.
Tip: Go through at least two rounds of improvements. The difference between your first and final prompt will surprise you.
AI & Ethics
Thinking about AI in society
Use this mode to explore topics like privacy, bias, deepfakes, AI and jobs, and who's responsible when AI gets things wrong. Spark won't tell you what to think — it asks questions and helps you build your own view.
Tip: Pick a topic you actually care about. The best discussions happen when you have a real opinion to defend.
Get more out of Spark
Spark won't do your work for you
That's the whole point. Every response will end with a question. Answer it — that's where the learning happens.
Be specific
Vague questions get vague answers. Instead of "help me with my code", try "my loop runs forever and I don't know why".
Use your voice
Tap the microphone button to speak instead of type. Spark will read your words — useful when you'd rather talk through a problem.
Don't just copy the answer
If Spark writes something, make sure you can explain it. Your teacher can see your conversations and will notice if you're just copying.
Try a different angle
If you're stuck, describe the problem differently. Explain what you've already tried. That usually unlocks things.
Ask "why"
If Spark says something you don't fully understand, ask it to explain. There are no silly questions here.
If anything in your Spark session makes you feel uncomfortable, worried, or unsafe — stop and tell your teacher straight away.
You can also contact Childline any time, for free, on 0800 1111. You do not have to give your name.
Log in and pick a mode. Your teacher will have given you your username and password.
Log in to Spark