❓What is Kita?

Kita is a platform for teaching Computing that lets teachers set programming work, assess it, and return feedback without moving between systems.

It starts by keeping the lesson whole. A lesson in Kita holds the instructions students read, the workspace where they write Python, the tests their programs must pass, the activity they submit to, the rubric it is marked against, and the feedback that comes back. Because all of that is attached to one lesson, the assessment cycle never leaves the platform — which is what makes fast, specific feedback practical for a full class.

Why this matters

Code editors work well because code is self-contained. Teaching is not.

Teaching is fragmented across systems. The editor holds the code, the VLE holds the task, a shared drive or inbox holds the submissions, and a spreadsheet holds the marks. Each system operates independently, with no shared understanding of how they connect.

Without that connection:

• A tool that only sees the code cannot tell whether the work was submitted, or submitted on time

• A tool that only holds the assignment cannot tell whether the program actually runs

• Feedback ends up separated from the code it refers to, so students read it without the work in front of them

• Every handoff between systems is manual, and the cost is paid per student, per lesson

Giving students a code editor is not enough. Assessment needs the task, the code, the tests and the mark scheme in the same place, tied to the same piece of work. That is what a Kita lesson is.

The lesson as the unit

When you build a lesson in Kita, everything the assessment cycle needs is attached to it:

• Instructions students read as they work, written and edited in the lesson editor

• A Python workspace in the browser, with no installation on school machines

• I/O tests — inputs and their expected outputs, which run when a student triggers a test run

• An assessed activity, if the lesson is one students submit to

• A rubric, with criteria, levels and points

• AI Help settings that apply to this lesson and no other

• Every submission, and the feedback returned against it

Because the lesson holds all of this, the work never has to be exported, renamed or re-uploaded. A student's submission arrives already attached to the task that set it, the tests that check it, and the rubric that scores it.

When a student's program produces an error, Kita shows the standard Python error output rather than a simplified substitute, so students practise reading the errors they will meet outside the classroom.

How AI supports students safely

Kita's AI support is called AI Help. It is off until a teacher turns it on, and it is configured for each lesson rather than for the platform.

This is the design decision that separates Kita from a general-purpose assistant. The teacher decides, lesson by lesson, how much help is available:

• Off by default. AI Help does nothing until a teacher enables it on a specific lesson.

• Each mode is independent. Explain, Debug, Tutor and Hints can be switched on or off separately, so a lesson can offer error explanations without offering hints.

• Hints escalate through capped levels. Support starts at the gentlest nudge and only increases if the student is still stuck. The teacher sets the maximum level the AI may reach.

• Tutor mode asks rather than answers. When a student expresses confusion, the Tutor agent responds with structured questions that prompt independent reasoning instead of supplying the solution.

• Feedback generation is teacher-triggered. The teacher chooses the style, length and format, and starts the generation. Students do not generate their own feedback.

The four hint levels are what make the cap meaningful:

Level 1 — Gentle Nudge. A broad reminder of relevant programming ideas, with no reference to the student's own code. Level 2 — Conceptual Hint. A pointer toward a specific concept, or a possible problem in their approach. Level 3 — Contextual Hint. A response that refers directly to the student's code and the logic they are implementing. Level 4 — Guided Next Step. An explicit instruction about the next change to make.

A lesson capped at L1 or L2 keeps the reasoning with the student. A lesson capped at L4 is closer to guided practice. The choice belongs to the teacher who knows the class.

This is not a chatbot with a content filter attached. The lesson settings and the AI's behaviour are the same system, so the limits are structural rather than advisory.

For the settings themselves, see Update the AI settings for a lesson.

Teaching and assessment

Alongside the workspace, Kita provides the structure teachers need to run a class and mark its work. Three concepts carry most of it: courses define what you teach, class groups define who you teach it to, and assessed activities define what gets marked.

Courses and lessons

A course is a container for related lessons. A lesson is a single unit of teaching — instructions, a workspace, and optionally an activity students submit to.

Lessons in the Time2Code format include an Evaluate section. An assessed activity usually sits there, after students have seen the expected outputs for the task.

Class groups

A class group is how students get into Kita. You create the group, enrol it onto the courses the class should see, and share a join link. Students use the link to add themselves, and the courses you enrolled the group onto appear for them.

Enrolment is managed at the group level, so adding a course for a class is one action rather than one per student.

See Create your class group and enrol students.

Assessed activities

An assessed activity is what makes work submittable and markable. Creating one lets you set a due date shown to students, the filenames you expect, how many I/O tests must be run or passed before submission, and the rubric the work is scored against. You can save it as a draft or publish it to students.

Expected filenames are a guide rather than a restriction: work submitted under a different name is still accepted, and a student who submits the wrong file can submit a replacement without waiting for the teacher.

See Create an assessed activity.

Marking and feedback

Submitted work arrives in Marking, already attached to the activity that set it. You open a student's submission, score it against your rubric, and return feedback.

There are three routes into a student's work, and which one you use depends on what you are doing:

• People — live teaching, watching the class code during the lesson

• Student progress — asynchronous checks between lessons

• Marking — working through submissions for an assessed activity

When AI generates feedback, three settings shape it, chosen independently of one another. Style is Concise, Supportive or Structured. Length is Shorter, Medium or Longer. Format is Standard prose or 2 Stars and a Wish. The defaults are Supportive, Medium and Standard.

If a student submits the wrong file, Unsubmit returns the work to them. Kita also accepts multiple submissions, so in most cases a student can correct the mistake without teacher intervention.

Kita for Schools and Open Kita

Kita for Schools is the licensed edition. Your school gets its own instance, with class groups, assessed activities, marking and administrative controls, governed by the Kita for Schools terms and privacy policy.

Open Kita is a free version for evaluating the platform without a school instance. Try it as a teacher or a student.

What Kita is not

• Not a general-purpose AI chatbot. AI Help is scoped to a lesson, and the teacher decides which modes are available and how far hints may go.

• Not a plagiarism detector. Kita does not judge whether submitted work is a student's own.

• Not a virtual learning environment. Kita covers Computing lessons, assessment and feedback. It does not replace your school's VLE or MIS.

• Not an installation. There is nothing for students to install. Kita runs in a browser.