Quadratic is an AI spreadsheet for data analysis, officially positioned as "the AI spreadsheet that connects to your data and helps you understand it quickly." It attempts to integrate natural language questions, traditional spreadsheets, code cells, and external data sources into a single collaborative, auditable workspace.
[Purpose] Quadratic is used to quickly turn data questions into explainable, shareable insights. The official description emphasizes that it says goodbye to =VLOOKUP, lets AI handle the tedious work, and delivers clear, correct, easy-to-understand results. It is suitable for individuals who need quick access to data and trustworthy answers, for small teams doing collaborative analysis, and for enterprises that need secure access to internal data and to apply AI.
[Main capabilities]
- AI spreadsheet and natural language questions: Users can ask questions in simple natural language, with formulas optional, to build spreadsheets faster. The official offering includes prompt suggestions.
- Multiple data source connections: Supports a variety of sources such as databases, spreadsheets, files, and APIs, with each source flowing into the same real-time spreadsheet, staying connected, synchronized, and available for AI to use. Official examples include connecting to Plaid to pull balances, transactions, and cash flow from more than 12,000 financial institutions, and pulling income statements, balance sheets, and invoice data directly from QuickBooks.
- Large data processing: The official description says it can open datasets that are too large for Excel or Google Sheets, and sorting, filtering, and AI analysis remain fast at the scale of 100 million rows.
- AI code cells: AI can write Python and SQL cells, allowing users to perform advanced analysis with code. The official description emphasizes that this is not a black box: users can open and edit any cell to inspect the AI's work and modify it as needed.
- MCP support: Quadratic supports the Model Context Protocol (MCP, an open standard for connecting AI with existing software). The official description says you can point Claude, Cursor, ChatGPT, or any MCP-compatible agent at Quadratic, letting the agent read, write, and verify work in the real-time spreadsheet, with every cell remaining auditable and every change approved by the user.
- Quadratic API: You can use your own code to trigger AI runs, read and write cells, and orchestrate agents, embedding Quadratic's spreadsheet intelligence into pipelines, proproducts and workflows.
- Collaboration and auditability: The official description says it supports team collaboration for verifiable analysis and keeps answers auditable.
- Security and compliance: Quadratic officially states that it has achieved SOC 2 and HIPAA certification; refer to its trust center.
[How to use] The official site provides an entry point to start for free (Get started for free). Users can connect their own data sources, ask natural language questions in the spreadsheet or use formula, Python, and SQL cells for analysis, and can invite teams to collaborate. Agents can connect via MCP, and developers can integrate via API. Specific installation commands, SDK usage, and detailed configuration steps are not yet provided in the official text.
[Use cases]
- Personal finance automation: Connect to Plaid to pull balances, transactions, and cash flow, and build auto-refreshing budget and net worth dashboards.
- Financial and accounting analysis: Retrieve income statements, balance sheets, and invoice data directly from QuickBooks without CSV export, and refresh at any time to get the latest financial data.
- Large dataset analysis: Open and analyze datasets beyond the handling capacity of Excel or Google Sheets.
- Team collaborative analysis: Collaborative, auditable analysis for small teams and enterprises.
- Agent-driven workflows: Let MCP-enabled AI agents read, write, and verify work in the spreadsheet, with users approving changes.
- Product and pipeline integration: Embed spreadsheet intelligence into your own systems via API.
[Limitations and notes] The available materials do not provide details such as specific pricing, free quota, platform support (such as Web, desktop, mobile), programming language versions, a complete list of data sources, or API rate limits. The official information only states SOC 2 and HIPAA certification and does not give more compliance details; the availability, regional restrictions, and configuration requirements of specific connectors (Plaid, QuickBooks, etc.) are also not yet provided.
