Home
The AI Toolkit dashboard — hero, Recommended tools, Recent activity, and the Build-a-student-chatbot card that assigns a guarded AI bot into an LMS course.
What this screen shows
The Home tab greets you (“Hi {name}, how can I help today?”) and gives three ways in: a Browse all tools button, a Recommended tools grid of nine everyday tools, and a Recent activity rail of your latest runs (click any to reopen that tool). Above the grid sits the Build a student chatbot card — the toolkit’s one route for handing an AI companion to your class safely.
On the dashboard
Jump to the All Tools catalog (all 125 tools, searchable and shelved by category).
A curated grid — Lesson Plan, MCQ Quiz, Worksheet, Rubric, Professional Email, Text Rewriter, Report-Card Comments, Concept Explainer and Reading Comprehension. Each card opens that tool’s runner.
Your most recent runs, newest first; click one to reopen the tool. Empty until you run your first tool.
Opens the chatbot builder — see the section below.
Build a student chatbot
The Build a student chatbot card opens a builder that assigns a course-grounded AI companion straight into an LMS course as a chatbot activity. Students chat with it inside the course — no toolkit permission required on their side. The builder is deliberately short: you name the bot, add optional instructions and a first message, then pick the course. The bot answers only from that course’s published content — the pages, books and files your students can already see — so it is grounded in exactly what you have taught, not the open web. This is the per-course face of Edukali’s knowledge-grounded bots; you manage what a bot knows (and re-sync it after you edit content) from Course Assistant settings in the course builder, and the same RAG engine powers the admin-built Knowledge Bots.
This replaces the older “persona” chatbot that you scoped by picking a bot type (tutor / chapter / study / character / custom) and a board-class-subject. Those pickers are gone — a bot no longer role-plays or works from the syllabus at large, it answers from the specific course’s materials, which is both safer and more accurate.
Grounded in course content
The bot answers from the target course’s published pages, books and files only — it will say it hasn’t been taught something rather than guess. What it knows is managed in Course Assistant settings (enable, auto-sync, re-sync after edits).
What students see the bot called (e.g. “Physics Course Assistant”). This becomes the activity’s title in the course.
Optional teaching rules (e.g. “ask one question back after explaining; use cricket examples; simple English”).
First message students see
Optional opening line to set the tone and invite the first question.
Pick a course and module; the bot lands as a chatbot activity students can start immediately.
Safety is non-negotiable and enforced server-side, always on. Guardrails cannot be turned off. A wellbeing layer watches conversations: a concerning message is flagged for you, and the bot itself shifts into a supportive, trusted-adult mode with helpline information.
You get a monitoring view of every bot: a list of student sessions (flagged ones surfaced first) and the full transcripts, so you can see exactly how your class is using it.