Practical overview
When it fits
Use LAMBDA when you want a conversational route from a dataset to code-backed exploration and reusable outputs such as charts, reports or notebooks.
Who it is for
- People who want to explore CSV, Excel, text and other common data files through natural-language questions.
- Analysts who want code-driven exploration, charts and reports gathered in a persistent conversation workspace.
- Teams or developers evaluating a self-hosted data-analysis agent with a configurable OpenAI-compatible model endpoint.
Documented uses
- Inspect a dataset and identify patterns by asking questions in ordinary language.
- Run Python or shell analysis code and create visualizations.
- Generate reports, notebooks and other artifacts from an analysis session.
- Use Autonomous Exploration to plan and carry out a hands-off dataset analysis.
Requirements & limits
- The repository documents Python 3.11+, Node.js 20+ and npm for local setup, plus an API key, compatible model endpoint and configured model list.
- The implementation identifies React and TypeScript for the frontend, FastAPI for the backend, and SQLite for local storage.
- LAMBDA runs analysis code on the backend machine; the project recommends trusted users and data and additional isolation before untrusted exposure.
- PDF and Chinese-report output needs optional TeX/XeLaTeX dependencies; Chinese output also requires CJK fonts.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Run LAMBDA locally
Configure the backend and start LAMBDA's local frontend and backend services.
Before you start
- Python 3.11 or newer
- Node.js 20 or newer and npm
- An API key and an OpenAI-compatible model endpoint with model IDs
Steps
Copy backend/.env.example to backend/.env.
Set OPENAI_API_KEY, OPENAI_BASE_URL and MODEL_LIST in backend/.env.
Run ./start.sh from the repository.
Open http://localhost:3000.
Expected outcome
The local LAMBDA app is served at http://localhost:3000.
Guide limitations
- Analysis code runs on the machine hosting the backend; the project advises using trusted users and data and adding isolation before exposing the backend to untrusted traffic.
- PDF and Chinese report generation requires optional TeX/XeLaTeX and CJK font dependencies.
Analyze a dataset and export session outputs
Use the app to explore a dataset and obtain analysis artifacts.
Before you start
- A running LAMBDA app
- A model selected from the model picker
- A data file to upload, or an example dataset
Steps
Choose a model in the app.
Upload one or more data files, or select an example dataset.
Ask a question about the data, or choose Autonomous Exploration for a planned analysis workflow.
Review generated artifacts in the Files panel and export available reports or notebooks.
Expected outcome
The conversation workspace can contain analysis results such as charts, tables, reports and notebooks; documented export formats include Jupyter notebooks, Markdown/report bundles, PDF reports and slides.
Guide limitations
- Analysis code executes on the backend machine.
- PDF and Chinese report support depends on optional system dependencies.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 18 October 2026
- Original website title
- AMA-CMFAI/LAMBDA · Official homepage title · retained in its source language