AIOTOR AILABS
Write real code
Use a browser IDE with backend Python intelligence, files, programs and contextual completion.
Launching 15 August 2026Python · ML · Deep Learning · Robotics
Executable intelligence learning
Learn the science.
Run the code.
See every state.
AIOTOR AILABS transforms difficult programming, machine-learning and robotics concepts into executable lessons. Learners change real code, inspect exact numerical state, watch scientifically labelled transitions and understand why the result changed.
80executable lessons
160practical use cases
4complete pathways
✓ Exact computed states✓ Browser-only output✓ Backend-authoritative execution
CONCEPTUAL MODELForward propagationSignals move through weighted connections
z = Σ(wᵢxᵢ) + bActivation: ReLUExact values appear after execution
PEDAGOGICAL TRANSITIONTraining and backpropagationForward pass → loss → gradient update
Loss: 0.084Epoch 12 / 18Interpolated movement is labelled, never presented as an extra training step
EXACT-STATE PREVIEWK-means clusteringAssignment and centroid recomputation
K = 3WCSS ↓ 286.42Centroids shown from the same numerical state used by the algorithm
SCIENTIFIC MODELProjectile motionPosition is computed from velocity, angle, gravity and time
x = v₀cos(θ)ty = v₀sin(θ)t − ½gt²Preview is illustrative; executed lessons show computed units and scale
A complete learning loop
From explanation to executable understanding.
Every lesson connects scientific foundations, editable source code, real output, runtime evidence and practical engineering use.
Execute safely
Rust orchestrates the learner workspace, validation, isolated execution and ownership.
Capture exact state
Variables, tensors, model metrics, coordinates, stdout and errors are recorded from the run.
See and explain
Frames, plots, spatial scenes, terminal output and narration stay synchronized in the browser.
Scientific integrity by design
Exact computed stateValues and coordinates come directly from the executed model.
Pedagogical transitionInterpolation is clearly marked between two valid computed states.
Conceptual modelAbstract mechanisms are labelled as explanatory representations.
The computation creates the truth. The visualization explains it.
AILABS never treats an attractive animation as scientific evidence. Every frame is classified and carries provenance, so learners know exactly what is computed, interpolated or conceptual.
{
"state_type": "exact",
"source": "python_execution",
"algorithm_step": "update_centroids",
"iteration": 4,
"validated": true,
"interpolated": false
}
Python execution→Typed state→Browser visual
Launch curriculum
Four connected pathways for intelligent-system builders.
Begin with Python, progress through object-oriented engineering and machine learning, then understand deep-learning systems through code and state.
01 from model import NeuralNetwork
02
03 network = NeuralNetwork([4, 8, 3])
04 history = network.fit(data, epochs=18)
05 network.predict(sample)
06 print(history.loss[-1])predict(sample)Jedi · project symbolPress Enter to insert
$ python train.py [epoch 01] loss 0.842 [epoch 06] loss 0.318 [epoch 12] loss 0.126 [epoch 18] loss 0.084 ✓ tests passed: 4 process exited successfully
One browser. A complete technical workspace.
Code, files, tests, terminal and visual output together.
- ✓Backend-aware IDE with Jedi semantic completion and project-file context.
- ✓Private learner folders for programs, modules, tests, documents and notes.
- ✓Browser-only scientific output from Matplotlib, Plotly, Bokeh and typed AIOTOR scenes.
- ✓Visible evidence through stdout, stderr, variables, tests and run history.
Simple launch access
One-time learning packages. No forced monthly subscription.
Pricing and GST are authoritative in the Rust backend. Package access activates only after verified payment confirmation.
No demonstration accounts or preset login credentials are included in this release.
Questions
Built to teach difficult ideas clearly.
AIOTOR AILABS combines real computation with carefully labelled visual explanation.
Is the visualization scientifically correct?
Executed lessons derive visuals from the same numerical state used by the code. Interpolated transitions and conceptual illustrations are labelled separately.
Does all output remain in the browser?
Yes. Terminal output, variables, SVG or PNG plots, Plotly figures, Bokeh figures, typed 2D scenes and completed spatial frames are delivered to authenticated browser panels.
Can learners change values and rerun the concept?
Yes. Validated controls update the program parameters, trigger backend execution and regenerate the state sequence and explanation.
Can administrators add more content?
Courses and lessons use validated content plug-ins. Administrators can create, edit, validate, reload and publish content without changing the learner frontend when an existing runtime family is used.
Are there any default accounts?
No. This release does not create demonstration users or ship preset passwords. Registration remains available, and a bootstrap administrator is created only when a secure password is explicitly configured.
AIOTOR AILABS
Build understanding that can be executed, inspected and trusted.
Create a learner account and begin with the free preview lesson in each pathway.
