Skip to main content
Tablix doesn’t use a single global “something crashed” screen. Instead, each subsystem (saving, AI, Python, SQL, sharing) surfaces its own errors close to where they happen — as an inline status, a browser alert, or a chat message — so you know exactly what failed and can retry that specific thing.
This page is a technical reference for what each error means and why it happens. For step-by-step fixes, see Troubleshooting.

Saving & Sync Errors

Autosave runs in the background: 5 seconds after your first edit to a new file, then every 2 minutes while the file has unsaved changes. A small sync status is tracked as idle → syncing → saved / error.
If saving fails because your Google session expired (an authentication error), reconnect Drive from Integrations and try saving again — your unsaved edits stay in the browser tab until you do.

AI Assistant Errors

Python Errors

Tablix’s Python engine (Pyodide) loads from a CDN the first time you open the code editor, showing “Loading Python runtime…” then “Initializing Python…”. If this never finishes, it usually means your network is blocking that CDN request.
The AI is specifically instructed to clean numeric columns and cast text columns before operating on them, and to avoid emojis in generated code — so these particular errors are more common in hand-written Python than AI-generated Python. See Python for the full bridge API.

SQL Errors

Chart Errors

Other In-App Messages

General Pattern

Across all of these, Tablix follows the same shape:
1

The failing action is isolated

A failed save, a failed chart render, or a failed AI call doesn’t take down the rest of the app — the grid, other sheets, and unrelated features keep working.
2

The message tells you what to do next

Messages point at a concrete next step: reconnect Drive, upgrade your plan, fix the code, or try again.
3

Retries happen automatically where it's safe to do so

AI calls retry on overload automatically; saves and chart renders do not auto-retry, so re-trigger them yourself if they fail.
For guided fixes to the most common issues, see Troubleshooting.