Chat with your data.
Answers you can reproduce.
An open-source, local-first conversational analytics workspace. A team of agents, powered by the model you prefer, queries your files and databases — then returns the SQL, charts, Python code, or notebooks behind every answer, ready to inspect and re-run.
macOS 13+ (Apple Silicon) · Windows 10+ (x64)
Linux or an Intel Mac? Install with Python →
Data sources
LLM Providers
From question to answer you can keep.
A team of agents
A coordinator routes each request to SQL, chart, notebook, and Python specialists. Every tool call stays visible, so you can check how an answer was produced.
Ground answers in Knowledge
Curate metrics, entities, joins, and golden queries as concepts the agent can cite, and review its proposals.
Connect your sources
Start with files and the managed lakehouse, then add PostgreSQL, MySQL, DuckDB, or SQLite.
Query your lakehouse
Imported files and generated artifacts are registered in a managed Iceberg lakehouse that is ready without credentials.
Explore the lakehouseChoose your model
Run local models with Ollama or LM Studio, or connect OpenAI, Anthropic, Gemini, and Bedrock.
See providersHow it works
Three steps from your data to analysis you can reopen and share.
Connect
Start with your files, the managed lakehouse, and the databases you already run.
Connect dataAnalyze
Ask in plain language and let the agent team write SQL, build charts, and draft notebooks.
See notebooksKeep
Save the SQL, charts, and notebooks the agents produce, with results stored in the document.
See outputsReady to ask your first question?
Install Kontext Studio, connect a source, and ask in plain language. Everything runs on your machine.
Latest Updates
Product changes, release notes, and practical guides.
Introducing Kontext Studio: An Open-Source, Local-First AI Data Workspace
Kontext Studio is now open source. Chat with your data in plain language — get the SQL, charts, Python code, or notebooks behind every answer, ready to inspect and re-run, on the model you prefer.