A course team staffed by agents.
Over 30 specialist agents read your policies, playbooks and recordings, draft the course, and check every line against the source. A named person approves it.
RafiqCourse BuilderHands over: course blueprint
Course Builder · arm or drop any tool
Click a tool to arm or disarm it; the trace re-runs live
- Cited to the sourceEvery generated line points back at your document.
- Your choice of modelRoute each stage to a provider you approve, or run it in your VPC.
- A person signs offNothing ships without a named approver.
- Arabic firstAuthored natively in Arabic. 30+ languages supported.
Your documents in, a checked course out.
The specialist agents draft from your material and check each line against it. A person approves before anything ships.
Why pricing rules existRafiq · drafting the lesson
QuizLina · writing the quiz
Fact-check passOmar · checking claims against the source
Owner of this blueprintReem · L&D leadAwaiting sign-offFive agents, five jobs. Try each one.
Queue a source, pick a clause, cut a module, hold the gate, then run the build.
Every source gets an id on arrival.
By the last stage, a quiz answer still points back to the file it came from.
Click a source to queue or remove it
Each id survives every downstream step; a quiz answer still points at pricing-policy-v4.pdf.
Reads it like a lawyer.
Clauses, tables, and exceptions come out anchored, so nothing floats free of its source.
Pick a clause; it holds while you read
Structure before a word of prose.
The plan arrives as modules you can cut. Click one to drop it; the run time and quiz bars follow.
No agent publishes itself.
A named owner signs the blueprint before a single lesson is written. Every decision is logged against their name.
You hold this one. Try it.
Reem · L&D leadOwner of this blueprintAwaiting sign-offModules land with receipts.
Writer, quiz, and fact-check agents work in parallel. Every claim carries the clause it came from.
Feed it anything with a lesson inside.
Policies live in PDFs. Expertise lives in recordings, decks, diagrams, and half-watched YouTube links. fkra reads all of it.
One module, four ways in.
fkra generates every module in all four modes at once, so it follows Universal Design for Learning from the first draft. Pick a mode to see what the learner gets.
The refund rule as a decision tree.
The refund rule is drawn as a branching path with the clause pinned to each branch. Color carries no meaning on its own.
Visual learners get the same governed content as everyone else.
4modes generated from the same cited source, no extra authoring pass.
Every output carries its source.
A reviewer can check any module, quiz, or answer against the clause it came from.
Every generated line points back at your document.
Claims the source does not support are flagged before a reviewer sees the draft.
Nothing ships without a named approver.
Your documents train no models. Retrieval only, isolated per workspace.
Bring your own model.
fkra does not tie you to one model. Route each stage to a provider you approve.
Click a provider chip to swap it. Routing is per stage, per workspace, and audit-logged.
Open-weights models, self-hosted. Documents and prompts never leave your network.
What security and IT teams ask about the AI.
Which models does fkra use, and can we choose?
You choose. Each stage (parse, plan, write, fact-check, translate) routes to a provider you approve, per workspace, and every routing change is audit-logged. Open-weights models can run in your VPC.
Do our documents train any model?
No. Your documents train no models. fkra uses them for retrieval only, and each workspace is isolated from the others.
How do we know the agents did not make something up?
Every generated line carries the clause it came from, so a reviewer can check it against your document. A fact-check pass flags claims the source does not support before a reviewer sees the draft.
Can an agent publish on its own?
No. A named owner signs the blueprint before a lesson is written, and every decision is logged against their name. Generation stays blocked until they approve.
Does it work in Arabic?
Arabic first. Courses are authored in Arabic with their own right-to-left layout, and the Translator agent keeps both versions in sync. 30+ languages are supported.
Can we keep the AI inside our own environment?
Yes. Enterprise deployments can run in your VPC or on-prem, with KSA data residency available. Talk to us about which fits your security review.
Bring one policy. See the draft.
We will show the agents drafting from your own document, every line cited, and a reviewer approving it.
Your documents train no models · Arabic and English