Distribution plots, correlation heatmaps, feature statistics — generated automatically for every column.
04
⬡
Model
AutoML · Compare
Train XGBoost, LightGBM, Random Forest side by side. Pick the best automatically. R², RMSE, accuracy all tracked.
05
→
Deploy
Reports · Endpoints
Export insights as PDF, serve predictions via REST endpoint, or re-run the whole pipeline on fresh data.
Live Demo
See it work in real time
dsagent — terminal
$
→
→
$
→
$
→
→
→
→
→
→
Capabilities
Everything built in
LLM
AI Analyst Chat
Ask questions about your data in plain English. Qwen with full tool-calling runs the actual analyses — no hallucinations, real results.
Builder
Pipeline Builder
Compose reusable workflows visually. AI suggests steps based on your dataset. Save, re-run on new data, view run history.
ML
AutoML
Select a target column. DSAgent detects regression vs classification, trains 3 models, compares them, and surfaces the winner.
XAI
Explainability
Feature importance, correlation heatmaps, and natural-language summaries baked into every analysis. No black boxes.
Persistence
Session Memory
Every upload, chat, and pipeline is persisted to PostgreSQL. Resume mid-analysis or reload a conversation weeks later.
Viz
Dark Charts
Histograms, scatter plots, correlation heatmaps, box plots — all styled in dark mode via matplotlib + imshow. PNG export.
Stack
Built on tools you already trust
Next.js 15, FastAPI, PostgreSQL via Prisma, Clerk auth — and Qwen-3 powering every intelligent decision.
Next.js 15
Frontend
FastAPI
Backend
Qwen
Intelligence
PostgreSQL
Database
XGBoost
ML Engine
LightGBM
ML Engine
Prisma
ORM
Clerk
Auth
FAQ
Frequently Asked Questions
DSAgent is an autonomous AI-powered data science platform that analyzes CSV datasets, cleans data, generates visualizations, trains machine learning models, and creates detailed reports automatically.
No. Simply upload your dataset and DSAgent handles data cleaning, exploratory analysis, feature engineering, model training, evaluation, and report generation without requiring coding knowledge.
Currently, DSAgent supports CSV datasets. After uploading, the AI automatically extracts metadata, analyzes your data, and provides insights through interactive visualizations and machine learning pipelines.
Yes. Along with the one-click autonomous workflow, DSAgent includes a drag-and-drop Pipeline Builder where you can customize every step, configure tools, and save reusable workflows.
Absolutely. DSAgent automatically saves the best-performing machine learning model, lets you download it as a deployment-ready bundle, and generates PDF reports that can also be emailed directly to you.
Get Started
Start in seconds
No code. No configuration. Upload your dataset and let DSAgent handle the rest.