FEATURES

QuerylessAI for Data Portals: Features

QuerylessAI adds an AI interface on top of a PortalJS data portal. Its main value is not generic chat, but grounded interaction with the portal, its datasets, and the data files behind them.

Grounded answers, not generic chat

Responses are backed by API calls and data queries

Queryless has tools to query APIs and data files directly. It is not meant to hallucinate answers or mix in unsupported sources when answering portal-related questions.

Answers can be grounded in the underlying data

When a user asks a question about a dataset or resource, Queryless can download and query the actual data instead of guessing based on partial inspection.

This matters because some LLM-based systems inspect only a fragment of a file and then generate answers from incomplete context. Queryless is designed to reduce that failure mode by explicitly querying the data.

Cross-dataset analysis

When multiple datasets cover a common topic, Queryless can query more than one dataset and combine the results into a single answer or output.

Context-aware portal assistance

Queryless can preserve conversation context while the user moves across different pages in the portal. This makes the experience more continuous than AI widgets that reset whenever the user changes page.

The agent knows what the user is currently looking at

Queryless receives context about the current page, which helps it interpret user intent more naturally.

For example, if a user is already on a dataset page and asks "show me a sample of this data", the agent does not need to ask which dataset they mean.

The agent can help users find the right content in the portal

Queryless understands how the portal is structured and can direct users to the relevant dataset, resource, or page.

This makes it useful not only for analysis, but also for search, discovery, and portal navigation.

Analysis and output capabilities

Data querying for question answering

Queryless can query dataset files to answer concrete user questions rather than relying only on metadata or text descriptions.

Visualizations powered by Vega

Queryless can generate insights and visualizations using Vega, which gives it access to a broad range of charting capabilities.

Shareable reports powered by Observable Framework

Queryless can create dashboards or reports with multiple visualizations, each available at its own URL.

These reports can be refined through the chat and then shared with other users.

Export of filtered or generated outputs

If a user wants to keep working with the result of an analysis, Queryless can export data in formats such as CSV or JSON.

This can apply to filtered outputs, derived result sets, or data shown in a visualization.

Accessibility and language

Multilingual interaction

Although product messaging may say "Ask questions in plain English", Queryless is not limited to English and can work in other languages as well.