ZenQuery Documentation

Table of Contents

Usage Guide

  • Click the 📁+ icon to create a new session and give it a descriptive name (e.g., "Sales Analysis Q4 2024").
  • Select your newly created session from the sessions dropdown menu and click 'Load Selected Session'.
  • Click on "Add Files" to upload files for analysis. ZenQuery supports following file types: csv, json, excel, ndjson, jsonl, parquet.
  • Ensure your LLM configuration is set correctly in settings before proceeding to next step (see Getting Your LLM API Key and Configuring Your Key in ZenQuery sections below).
  • That's it!
  • Start asking questions in natural language in the query input box. Examples:
    • "What are the top 10 products by sales volume?"
    • "Show me email campaigns with the highest open rates"
    • "Who are our top 20 customers by revenue in 2024?"
  • You can create multiple sessions with different sets of files and switch between them quickly whenever needed.
  • Pro Tip: Use the query context box (top right) for additional query generation context like formatting or ordering preferences. This context is saved per session and applies to all queries. Example: "Format all decimal numbers to a max of 2 decimal places and sort results by date ascending"
  • Important: While LLMs are powerful, they can occasionally make mistakes. Always double-check the generated output before using it for critical decisions.

Getting Your LLM API Key

ZenQuery requires an API key from a compatible LLM provider. We recommend OpenRouter for its flexibility and cost-effectiveness, but OpenAI is also fully supported. Any provider with OpenAI compatibility (>99%) will work.

Using OpenRouter (Recommended)

OpenRouter provides access to multiple AI models at competitive prices, making it ideal for most users.

  • Visit openrouter.ai and create a new account.
  • Add at least $5 in credits to your account. This amount supports tens of thousands of queries.
  • Navigate to your API Keys page.
  • Click "Create Key", enter a descriptive name, and click "Create".
  • Copy the generated key (it starts with sk-or-v1-...). Store this key securely.

Using OpenAI

If you prefer to use OpenAI directly or already have an account, you can use their API.

Configuring Your Key in ZenQuery

Once you have your API key, follow these steps to set it up inside ZenQuery.

  • Open ZenQuery and click the settings button (gear icon) in the bottom left corner.
  • In the settings panel, paste your API key into the "API Key" field.
  • Set the "Endpoint" field based on your provider:
    • For OpenRouter: https://openrouter.ai/api/v1
    • For OpenAI: https://api.openai.com/v1
  • In the model field, we recommend starting with openai/gpt-4o for the best results, then switching to openai/gpt-4.1-mini for incredible cost savings with almost identical quality for daily usage.
  • Click "Save" to store your configuration.
🔒 Security Note

Your API key is stored securely and encrypted on your local device. It is never sent to ZenQuery servers and is only used to communicate directly with your chosen LLM provider.

ZenQuery Video Guides

🚀 Quickstart (Under 2 Minutes)

📘 Detailed Usage Guide

Understanding Running Costs

ZenQuery is designed to be extremely cost-effective. App is one-time payment, and LLM costs associated with usage are negligible. See below.

Recommended Models & Costs

Below are our recommended models with estimated questions per dollar, based on a single CSV file with 10 numeric columns.

Model Name Quality Questions Per Dollar Best For
openai/gpt-4.1-mini ⭐ Close to Best ~4,500 Daily usage for simple to complex questions
openai/gpt-4o Best ~300 Very complex and detailed questions
google/gemini-2.0-flash Very Good ~18,000 Extremely heavy usage up to moderately complex questions
deepseek/deepseek-chat-v3 Very Good ~6,500 Heavy usage up to moderately complex questions
openai/gpt-4o-mini Good ~12,000 Very heavy usage up to low-moderately complex questions
💡 Cost Example

With our recommended openai/gpt-4.1-mini model, a $5 credit from your provider could answer roughly 25,000 questions of yours — enough for a complete year of regular data analysis! And remember, the size of files does NOT matter at all.

Data Privacy & Security

Your privacy is paramount. By default, only the first 3 rows of your files are shared with the LLM to provide context for analysis. You have full control and can adjust this number down to zero in the settings panel. Your files are never uploaded on any server.

Need Help?

If you have questions not covered here or need assistance, we're ready to help.