data-formulator
SHA-256An open-source AI workspace from Microsoft Research that turns natural language questions into editable charts, with branching analysis threads and support for multiple databases and LLM providers.
Smart Download
Download Download Version
v0.8b1 · 255.8 MB
Chat with your data to create charts — open-source and model-agnostic.
Core Features
- Natural language to chart: ask questions in plain English and get 30+ editable chart types
- Data Threads: branch from any step to compare alternative analyses without losing context
- Broad connectivity: files, screenshots, text, MySQL, PostgreSQL, MSSQL, Databricks, S3, Azure Blob and more
- Multi-LLM support via LiteLLM: OpenAI, Azure, Anthropic, or fully local with Ollama
- Data extraction from images and unstructured text
What It Can't Do
- •You must supply your own LLM API key — cloud models incur usage costs and your data is sent to the provider, so use a local Ollama model for sensitive data. Desktop preview builds are not code-signed, so Windows SmartScreen and macOS Gatekeeper will show warnings you must manually bypass. AI-generated insights can be wrong, so verify results against the raw data before making decisions. It is an exploratory analysis tool, not a full enterprise BI replacement: no role-based access control, scheduled refresh, or multi-user collaboration like Tableau or Power BI.
Use Cases
- Business users exploring spreadsheets without writing any code
- Analysts rapidly prototyping visualizations for reports and presentations
- Querying live databases in plain language to monitor metrics
- Turning data tables from screenshots or web pages into interactive charts
Data Formulator is an open-source, AI-powered data exploration tool from Microsoft Research that turns natural language questions into editable, polished visualizations. Upload CSV, Excel, JSON, screenshots, or plain text — or connect live sources like MySQL, PostgreSQL, MSSQL, Azure Data Explorer, S3, and Databricks — then chat with an AI analyst to generate 30+ chart types powered by the Flint visualization engine. Its signature Data Thread feature lets you branch from any step, compare alternative analyses, and keep full context, solving the lost-in-chat-history problem common to AI analytics tools. Through LiteLLM it works with OpenAI, Azure, Anthropic, and local Ollama models, so you can run it fully offline. Compared with ChatGPT Advanced Data Analysis or Tableau, Data Formulator is free, open-source (MIT), model-agnostic, and offers branching exploration instead of a single linear conversation. Install in one command via uvx or pip, run with Docker, or download desktop builds for Windows and macOS.
Tags
Getting Started
Download installer
Click the button above to download the installer for your system
Install the software
Open the downloaded dmg file, then drag the app to Applications
Fastest option: try the hosted demo at data-formulator.ai with no installation
For local use, run uvx data_formulator in a terminal, or pip install data_formulator followed by python -m data_formulator
Open the automatically launched page at http://localhost:5567 in your browser
- Fastest option: try the hosted demo at data-formulator.ai with no installation
- For local use, run uvx data_formulator in a terminal, or pip install data_formulator followed by python -m data_formulator
- Open the automatically launched page at http://localhost:5567 in your browser
SHA-256 checksum verified
Checksum extracted from GitHub official Release page
SHA256 Checksum
78b2bab08a9a6260c4e8255cd2baaa6946d6599d398bd55559f3d1e3ac084ad7This checksum is extracted from the GitHub Release page. Verify file integrity after download.
All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.
Open Source Transparency
View GitHub SourceUninstall Info
If installed via pip, run pip uninstall data_formulator. If launched with uvx, run uv cache clean to remove cached files. For Docker, run docker compose down and delete the images. For desktop builds, simply delete the app or executable. Local user data is stored in a local directory and can be removed manually after uninstalling.
No Extra Dependencies
Ready to use after download. No additional runtime required.
Having issues? Check the FAQ below
3 FAQs
Similar Projects
ollama
Ollama lets you download, run, and manage large language models locally. One command, multiple platforms, endless possibilities.
llama.cpp
High-performance LLM inference engine in C/C++ with minimal dependencies, supporting quantized models (1.5–8 bit) and diverse hardware (Apple Silicon, CUDA, Vulkan, etc.).
mempalace
Open-source local-first AI memory system with verbatim storage and semantic search — zero API calls, no keys, top benchmark scores.