Data Talks

    Talk to your data, get answers.

    The open-source AI workspace for data analysis. Connect CSV, BigQuery, Google Sheets or SQL databases, ask questions in natural language, and get visual answers with charts and tables.

    Get StartedRead the docs
    The pillars

    One workspace. Four ways to understand your data.

    Data Talks unifies data connection, conversational AI and visual analytics. Your team asks questions in plain language and gets instant answers — no SQL, no dashboards to build.

    35+ data sources

    CSV, XLSX, BigQuery, Google Sheets, PostgreSQL, MySQL, MongoDB, Salesforce, HubSpot, Stripe, and more.

    Learn more

    Natural language Q&A

    Ask questions in Portuguese, English or Spanish. The AI agent generates queries, runs them, and explains the results.

    Learn more

    Charts & dashboards

    Auto-generated charts from every conversation. Compose them into custom dashboards and share with your team.

    Learn more

    Recurring alerts

    Set daily, weekly or monthly alerts. Get notified when your data changes — no manual monitoring.

    Learn more
    Everything included

    The whole toolkit, in one workspace.

    All the features you need to turn raw data into actionable insights.

    CSV / XLSXBigQueryGoogle SheetsSQL databasesNatural language queriesAuto chartsAlertsDashboardsMulti-languageStudio reportsAudio narrationAPI keysTelegram / WhatsAppCDP workspaceETL pipelinesAudit trailDark modeSelf-hosted
    Built on
    ReactViteTailwindshadcn/uiLangFlowPythonFastAPI
    FAQ

    Frequently asked questions

    What is Data Talks?

    Data Talks is an open-source AI workspace for talking to your data in natural language. Connect a CSV, BigQuery, Google Sheets, Postgres or any SQL database, ask questions, and get answers with charts.

    Is it really free?

    Yes. Data Talks is open source under the Apache 2.0 license. You can run it for free on your own infrastructure with Docker. The only cost is the LLM tokens you choose to spend.

    Which data sources are supported?

    Over 35 sources including BigQuery, PostgreSQL, MySQL, SQL Server, Snowflake, Redshift, SQLite, Google Sheets, CSV, XLSX, Salesforce, HubSpot, Stripe, GA4, and more.

    Can I self-host it?

    Yes. Data Talks is designed self-hosted-first. Spin it up with Docker Compose on your own VPS or Kubernetes cluster — no telemetry, no usage limits.

    How does it compare to Metabase, Wren AI or Chat2DB?

    We maintain an honest comparison page. The short version: Data Talks wins on multilingual UX (PT/EN/ES), recurring NL alerts, and Google Sheets connectivity. The others win on community size and (for Wren AI) semantic-layer maturity.

    Which LLM does it use?

    Pluggable. OpenAI, Anthropic, Ollama, and any LiteLLM-compatible model. Configure it per workspace — useful for data residency or local inference.

    Can I use it in production?

    Yes. Pin charts to dashboards, schedule recurring alerts via webhook or Telegram, and inspect the raw SQL behind every answer for auditability.