# DashboardGPT > DashboardGPT is an AI-powered visual BI workspace that lets users upload data, ask questions using natural language, generate analytics, and organize charts, KPIs, and insights on an interactive canvas. DashboardGPT, a web application operated by Andolasoft Inc., lets you bring a dataset (CSV or Excel upload; warehouse connections on paid plans are coming soon) and ask questions in plain language; each answer lands on an infinite canvas as a chart, a metric, a table or a written insight. Follow-up questions build on what is already on the canvas. Query execution runs on an Apache Superset-powered analytics layer, and customers can bring their own AI provider key. Key capabilities: AI analytics, Infinite canvas, Natural language, Bring your own data, Bring your own key, Written insights, Semantic understanding, Superset-powered analytics. Plans: Explore (Free), Team ($40 per user / month), Enterprise (Custom). A public playground at https://www.dashboardgpt.org/playground runs the product on a demo dataset with no account. An on-premise edition installs with Docker Compose or Kubernetes, including air-gapped environments with a local LLM; it is licensed annually, from $9,000 per year for Self-Hosted Team and $24,000 per year for Self-Hosted Business, with Enterprise Air-Gapped quoted individually. Every page listed below is also available as HTML at the same URL without the `.md` suffix, and any page returns Markdown when requested with `Accept: text/markdown`. ## Product - [DashboardGPT home](https://www.dashboardgpt.org/index.md): DashboardGPT turns natural-language questions into interactive charts, insights and visual analysis on an infinite canvas. - [Features](https://www.dashboardgpt.org/features.md): AI analytics, an infinite canvas, natural-language questions, bring-your-own data and bring-your-own key — everything in the DashboardGPT workspace. - [How it works](https://www.dashboardgpt.org/how-it-works.md): From a question in plain language to a chart on an infinite canvas: how DashboardGPT resolves intent, runs the query and chooses the visualisation. - [Pricing](https://www.dashboardgpt.org/pricing.md): Start free with CSV and Excel uploads. Add shared canvases, higher limits and access control on the Team plan, with warehouse connections coming soon, or talk to us about Enterprise. - [Security](https://www.dashboardgpt.org/security.md): Tenant isolation, customer-owned AI keys, encrypted data processing and workspace access control — how DashboardGPT handles your data. - [ChatGPT alternative for data analysis](https://www.dashboardgpt.org/chatgpt-alternative-for-data-analysis.md): A ChatGPT alternative for data analysis: ask questions about CSV and Excel files in plain English and keep every chart and insight on a shared canvas, with the AI provider you choose. - [FAQ](https://www.dashboardgpt.org/faq.md): Common questions about DashboardGPT: what it is, how the playground works, which data sources are supported, and how AI credentials are handled. - [What's new](https://www.dashboardgpt.org/whats-new.md): New features, pricing updates and improvements shipped to DashboardGPT, newest first: sharper charts, clearer answers and simpler plans. ## On-premise (self-hosted) - [On-Premise](https://www.dashboardgpt.org/on-premise.md): Run DashboardGPT on your own servers with Docker Compose or Kubernetes. Bring your own LLM key or run a local model fully air-gapped. Your data and prompts stay inside your network. - [System requirements](https://www.dashboardgpt.org/on-premise/system-requirements.md): Hardware, software, database, network and GPU requirements for running DashboardGPT on-premise with Docker Compose, Kubernetes or a fully air-gapped install. - [On-prem pricing](https://www.dashboardgpt.org/on-premise/pricing.md): Self-hosted DashboardGPT plans for teams, businesses and air-gapped enterprises. Annual licenses with updates, support and install help. Book a call for a quote. ## Product experience - [Playground](https://www.dashboardgpt.org/playground.md): Try DashboardGPT on a live demo dataset: ask questions in plain language and watch charts, metrics and insights appear on an infinite canvas. ## Learn - [Learn](https://www.dashboardgpt.org/learn.md): Guides, comparisons and templates for conversational analytics: chat with your data, BI alternatives, Apache Superset + AI, analytics templates and bring-your-own-key privacy. - [Chat with your data: the complete guide to AI data analysis](https://www.dashboardgpt.org/learn/chat-with-your-data.md): How to chat with a CSV, Excel file or database in plain English and get charts, metrics and written insights back. A practical guide with example questions. - [The best conversational BI tools in 2026, compared](https://www.dashboardgpt.org/learn/best-conversational-bi-tools.md): Compare conversational BI and AI data analysis tools: chat-first assistants, enterprise BI with AI, notebooks and canvas workspaces. Pick the right fit for your team. - [Apache Superset + AI: adding natural language to your BI stack](https://www.dashboardgpt.org/learn/apache-superset-ai.md): How to add natural-language questions and AI insights to Apache Superset. Text-to-SQL on Superset, semantic profiling, guardrails and when to use a canvas. - [Analytics templates: the questions to ask your business data](https://www.dashboardgpt.org/learn/analytics-templates.md): Ready-to-use question sets for SaaS metrics, marketing spend, sales pipeline, finance and operations. Ask them in plain English and get charts in seconds. - [Bring your own key: private AI analytics, explained](https://www.dashboardgpt.org/learn/bring-your-own-key.md): What bring-your-own-key means for AI analytics: who bills for model usage, what data reaches the AI provider, how tenant isolation works and how to set it up. - [How to chat with a CSV file and get charts, not paragraphs](https://www.dashboardgpt.org/learn/how-to-chat-with-a-csv-file.md): A step-by-step walkthrough for asking questions of a CSV in plain English and getting charts back. Prepare the file, upload it, ask, follow up and check the numbers. - [Natural language to chart: how AI turns a question into a visual](https://www.dashboardgpt.org/learn/natural-language-to-chart.md): What happens between typing a question and seeing a chart: parsing intent, mapping to columns, writing the query, choosing the chart and writing the insight. - [Excel pivot tables vs asking a question: a side-by-side test](https://www.dashboardgpt.org/learn/excel-pivot-tables-vs-ai.md): The same five business questions answered with Excel pivot tables and with plain-English questions. Steps, trade-offs and when to use each approach. - [How to test an AI data analyst: 50 questions and the ways answers go wrong](https://www.dashboardgpt.org/learn/50-questions-ai-accuracy.md): A practical test kit for checking any AI data analysis tool: 50 questions in five difficulty levels, the eight common failure types and how to score the results. - [DashboardGPT vs Julius AI: canvas or chat thread?](https://www.dashboardgpt.org/learn/dashboardgpt-vs-julius-ai.md): DashboardGPT vs Julius AI compared: how answers are kept, follow-up questions, data sources, AI keys and pricing. An honest guide to which tool fits which team. - [ThoughtSpot and Power BI Copilot alternatives for teams under 200](https://www.dashboardgpt.org/learn/thoughtspot-power-bi-copilot-alternatives.md): Enterprise AI analytics without enterprise setup. When ThoughtSpot or Power BI Copilot fit, when they don't, and what smaller teams use instead. - [ChatGPT data analysis vs DashboardGPT: when each one wins](https://www.dashboardgpt.org/learn/chatgpt-data-analysis-vs-dashboardgpt.md): ChatGPT is great for one-off file analysis. DashboardGPT is built for analyses you keep, share and build on. A practical guide to when each one wins. - [Free Tableau alternatives for teams that don't write SQL](https://www.dashboardgpt.org/learn/free-tableau-alternatives.md): Six free or free-tier alternatives to Tableau for non-technical teams, compared on how you build charts, what data they take and the catch in each free option. - [Adding natural-language questions to Apache Superset](https://www.dashboardgpt.org/learn/superset-natural-language-questions.md): A technical guide to adding natural-language questions to Apache Superset: schema context, prompt design, SQL validation, execution through Superset and chart selection. - [Superset text-to-SQL: what works and what breaks](https://www.dashboardgpt.org/learn/superset-text-to-sql.md): Where text-to-SQL on Apache Superset works well, the seven query patterns that break it (joins, date grains, ambiguity, NULLs) and a fix for each. - [Semantic profiling: how the AI learns what your columns mean](https://www.dashboardgpt.org/learn/semantic-column-profiling.md): How semantic profiling detects column types, roles, units and relationships before the first question, and how to help it with better column names. - [From Superset dashboards to an infinite canvas](https://www.dashboardgpt.org/learn/superset-dashboards-to-canvas.md): Why a grid of dashboard tiles struggles with follow-up questions, what an infinite canvas does differently, and how to use both on a Superset stack. - [5 questions every SaaS founder should ask their Stripe export](https://www.dashboardgpt.org/learn/saas-founder-stripe-questions.md): Turn a Stripe subscriptions export into answers on MRR, churn, expansion, cohort retention and revenue concentration. Five questions, the columns you need and what to look for. - [Marketing spend analysis in 3 questions](https://www.dashboardgpt.org/learn/marketing-spend-analysis.md): Three plain-English questions that find the marketing channel quietly eating your budget: cost per signup by channel, trend by week and campaigns to cut. - [Sales pipeline analysis without SQL](https://www.dashboardgpt.org/learn/sales-pipeline-analysis-without-sql.md): Analyse a CRM deals export in plain English: pipeline by stage, stage conversion, win rate, cycle time and rep performance. Questions, columns and what to look for. - [Why your analytics tool shouldn't mark up AI tokens](https://www.dashboardgpt.org/learn/why-byok-analytics.md): Bundled AI pricing hides the real cost of model usage. The case for bring-your-own-key analytics: transparent cost, model choice and control of your data agreement. - [How to get an OpenAI or Anthropic API key in 2 minutes](https://www.dashboardgpt.org/learn/get-ai-api-key.md): Step-by-step: create an OpenAI or Anthropic API key, set a monthly spend limit, keep it secure and connect it to your DashboardGPT workspace. - [Is my data used to train models? A plain-English answer](https://www.dashboardgpt.org/learn/is-my-data-used-to-train-models.md): What actually reaches an AI provider when you use AI analytics, what major providers say about training on API data, and the questions to ask any vendor. - [Your Apache Superset dashboards, right inside your workspace](https://www.dashboardgpt.org/learn/embed-superset-dashboards.md): See your Apache Superset dashboards and charts right inside your DashboardGPT workspace. No extra logins for your team, and each workspace only ever sees its own data. - [From AI canvas to Superset dashboard in one click](https://www.dashboardgpt.org/learn/publish-canvas-to-superset.md): Ask questions in plain English, arrange the answers on a canvas, and publish it as a shareable Apache Superset dashboard that looks just like your canvas. - [Why the canvas chart and the Superset chart always match](https://www.dashboardgpt.org/learn/canvas-and-superset-charts-match.md): Stop arguing about whose number is right. With DashboardGPT, the chart on your canvas and the chart in Apache Superset always show the same numbers. - [Per-user BI pricing: choosing Team seats, monthly vs annual](https://www.dashboardgpt.org/learn/per-user-bi-pricing.md): Simple, predictable per-user BI pricing: how to choose the right number of users, what annual billing saves, and what DashboardGPT Team costs for 5, 10, 25 and 50 users. - [No surprise AI bills: tracking AI usage and cost in your workspace](https://www.dashboardgpt.org/learn/track-ai-usage-and-cost.md): See how much your team uses AI at a glance, pay your AI provider directly with your own key, and get bonus runs when you sign up. No surprise bills. - [Run DashboardGPT on your own AI model: Ollama, vLLM and LM Studio](https://www.dashboardgpt.org/learn/local-models-openai-compatible.md): Keep your data in-house by running DashboardGPT on the AI model of your choice, including models on your own servers such as Ollama, vLLM or LM Studio. - [Self-hosting DashboardGPT: an evaluation checklist for your team](https://www.dashboardgpt.org/learn/self-hosted-analytics-checklist.md): A simple evaluation checklist for teams that want DashboardGPT on their own servers: how to run it, which AI to use, who gets access, and what it costs. - [How to run a team trial of an AI analytics tool (and decide with confidence)](https://www.dashboardgpt.org/learn/team-trial-checklist.md): A practical checklist for evaluating an AI analytics tool with your team: who to invite, which questions to test on real data, how to judge the answers, and what to do when the trial ends. - [Your free Team trial: what you get, and how to choose what comes next](https://www.dashboardgpt.org/learn/free-team-trial.md): Every new DashboardGPT account starts on a free Team trial with no card. See what the trial includes, how to get the most from it, and how to choose between the free Explore plan and Team when it ends. ## Company / Contact - [Contact](https://www.dashboardgpt.org/contact.md): Talk to the DashboardGPT team about Enterprise deployment, security review, or whether the product fits how your team works. ## Optional - [Terms of Service](https://www.dashboardgpt.org/terms.md): The agreement between you and Andolasoft Inc. that governs your use of DashboardGPT. - [Privacy Policy](https://www.dashboardgpt.org/privacy.md): How Andolasoft Inc. collects, uses and protects personal data when you use DashboardGPT.