How is Analytics Agent different from Grafana?
Grafana focuses on observability — watching system metrics. Analytics Agent is about interactive visual analytics: digging deep into data, not just indicator panels. Reports are edited by the AI in plain language, Jupyter Notebooks embed right into them, documents are JSON files made for Git, and three connection strings cover the data sources. The product was created by an analyst with more than 25 years of experience working with data.
How does the AI add reports and charts?
Next to any block or error there is an AI button: describe in words what you need (“add a weekly sales chart”, “make a table of the top 10 customers”), pick a model — and an agent chat opens. The agent edits the report’s JSON directly, the file watcher picks up the change, and the chart appears on screen instantly.
What are analysis templates?
It is a ready-made library of 257 data analysis solutions: sales forecasts, customer segments, assortment, advertising, Wildberries and Ozon reports — plus a free 50-lesson analytics course. Open a template, click “Use this template” — and the agent adapts it to your data, runs the calculations and installs the finished report into Analytics Agent. Details in the Analysis Templates section.
Which charts can I build?
28 chart types — from lines, bars and scatter plots to funnels, Gantt charts, flows, region maps and relationship graphs. And every chart shares the same capabilities: reference lines and normal ranges, intervals, value labels, comparison with the previous period, and axis settings. The full list with advice on when to use which is in the Charts section.
Which databases are supported?
SQLite, PostgreSQL and ClickHouse as full data sources with schema auto-detection. Plus the Lakehouse: queries to Parquet and SQLite files right from the file browser, without configuring any sources.
Do I need to know how to program?
To read reports — no. SQL helps for your own queries, but the AI will write the query for you from a verbal description. For heavy analytics there is a Jupyter Notebook with Python right in the report (pandas, NumPy, Matplotlib, Seaborn are all included).
What do I need to run it?
A Docker host: one container, a volume for the data, a login and the SHA-256 hash of a password in environment variables. AI features turn on by pointing to the URL and API key of your Xedant Agent.
Can I use it without Xedant Agent?
Yes: reports, charts, tables, data sources, validation, Jupyter, the Lakehouse and analysis templates all work on their own. The integration adds the AI buttons — adding and fixing blocks in plain language, and adapting templates to your data.
Where are reports stored?
As JSON files in the /data/documents volume — you can version them in Git, edit them by hand, and move them by copying. Changes are picked up instantly, without a reload.
Can I send a report to someone without access?
Yes, in two ways. A report can be exported to PDF — a ready file for an email or a meeting. And if a live report is what you need, a document can be made public and shown by link, without sign-in. Details in the Reports & Documents section.