FORGE

AI-driven scikit-learn playground and compiler. Describe your model to generate it, or write and run your own ML code directly in the browser without any setup.

01

Describe

Type what you want in plain English. No setup, no IDE, no blank notebook. The AI understands intent, picks the right algorithm, and starts writing immediately.

02

Watch it build

The agent writes production Python live, character by character, on screen. You watch the model, the plots, and the parameters take shape in real time before anything runs.

03

Run instantly

Your code executes in your browser — no server, no install, no waiting. Powered by Pyodide, sklearn runs client-side. Metrics and live plots appear the moment training finishes.

04

Interact

Every parameter becomes a slider. Change tree depth, swap the criterion, toggle the decision boundary. The model retrains in under a second — data stays loaded, only training reruns.

Why Forge

The gap is zero.

Other tools give you code in a text box. You copy it, set up an environment, debug it, run it yourself. Forge gives you a running, interactive experiment. The distance between "AI wrote it" and "it's live" is zero.

No environment. Ever.

Python runs entirely in your browser tab via WebAssembly. scikit-learn, numpy, pandas, matplotlib — all available instantly, zero install, zero config, zero cost to you.

The AI designs its own UI

The agent doesn't just write the model — it declares every slider, dropdown, and toggle in a schema. Your controls panel is built automatically from what the AI decided to expose.

See only what matters

Flask boilerplate, routing, base64 encoding — all hidden. The optional code drawer shows only the sklearn section. Beginners see results. Power users see the model logic.

Built-in ML Compiler

Want to write code yourself? Use our built-in Compiler to write and run your custom ML code directly in the browser without any setup. Test models instantly.

Every algorithm. One interface.

Decision trees, random forests, k-means, SVMs, PCA, regression — and anything else you can describe. The agent writes fresh code for every request. No templates. No limits.

Upload your own data

Drop a CSV and the agent builds the entire pipeline around it — detects column types, suggests the target variable, wires up preprocessing. Your data, your model, in seconds.

05

Collaborate

Share sessions with one click. Teammates can fork models, tweak parameters, and compare results in their own secure sandbox. No setup required.

06

Export

One click to production. Export your final model as a trained Pickle file or a scikit-learn script ready for use in any Python environment.

07

Privacy

Forge runs entirely in your browser. scikit-learn training stays local—none of your raw CSV data ever touches our servers. Your data stays yours.

08

Mastery

Learn by doing. Hover over any parameter or metric to see plain-English explanations of what it is and how it affects your model's performance.

Community

Playgrounds worth sharing.

Every Forge session can become a public playground. Fork, rate, embed. The best experiments surface to the top.

fork

Build on anyone's session

Open a shared playground, change the parameters, and save it as your own starting point. Their model architecture, your experiment. No attribution lost.

rate

Community-graded quality

Rate any playground on clarity and usefulness. The highest-rated experiments surface in search and the featured gallery — no algorithmic noise, just what's actually good.

embed

Live models in any page

Every playground has an embed link. Drop it into a blog post, a course, a README. Readers interact with the live running model without leaving the page.

Pricing

Simple, transparent pricing

Start free. Upgrade when you need unlimited AI power.
Powered by Ethereum • Sepolia Testnet

Free

Get started with basic features

0ETH
  • Browser-based Python execution
  • Basic ML model generation
  • Community playgrounds access
  • CSV upload & preprocessing
  • 3 AI generations per day
  • Unlimited AI generations
  • Advanced model architectures
  • Priority execution queue
  • Export to production
Most Popular

Pro

Full power. No limits.

0.001ETH/month
  • Everything in Free
  • Unlimited AI generations
  • Advanced model architectures
  • Priority execution queue
  • Export to production
  • AI-powered code chat
  • Custom dataset training
  • Session history & restore

🔒 Payments are secured via Ethereum blockchain • Cancel anytime