> For the complete documentation index, see [llms.txt](https://docs.beye.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.beye.ai/release-notes/september-14-2026.md).

# September 14, 2026

This release adds charts to advanced analysis, lets a single analysis return several results at once, and fixes long answers that could hang without ever finishing.

#### New

**Charts from advanced analysis**

An analysis can now produce a chart directly, alongside the figures and tables it already returned.

Scatter plots are supported for the first time, including on numeric axes, so you can plot the relationship between two measures rather than forcing one of them onto categories.

Color points by a category, such as customer segment, region, or product line, to show how groups separate.

When the assistant clusters your data, it marks each cluster's center on the chart, so the segments it found are visible at a glance.

Charts appear in the conversation beside the table they were drawn from, and both classify each column the same way, so what you see plotted matches what you see listed.

#### Improved

**Several results from a single analysis**

An analysis previously returned one result. It can now return up to five.

Results appear in the order the assistant produced them, freely mixing summary figures, full tables, and charts.

A question such as "segment these customers and show me how the groups differ" can now come back as the segmentation chart, the per-segment summary, and the underlying table together, rather than as three separate questions.

#### Fixed

**Long answers that never finished**

Previously, a sufficiently long or detailed answer could fail to complete. The response sat showing a thinking indicator that never resolved, the conversation itself reported an error, and retrying produced the same result.

Long answers now always reach a finished state.

**Responses left hanging after an interruption**

Previously, the cleanup that closes out responses interrupted mid-answer could stall on large conversations and stop running entirely, leaving those responses stuck behind a thinking indicator.

Interrupted responses are now cleared reliably.