Maestrow Chatbot Upgrades
To provide the most advanced reasoning, speed, and accuracy, we periodically upgrade our core AI engine. While these upgrades bring significant improvements, they often require a fresh session state.
When we upgrade to a new model architecture, previous chat session data will be lost. Current chat histories cannot be migrated because the memory structure of the new model is incompatible with the old one.
Why is data cleared?
Think of an upgrade as a brain transplant. The new model processes information using different parameters and logic. To ensure the AI doesn't experience hallucinations or performance lag caused by outdated context formats, we start with a clean slate.
How to Save Your Work
Before an upgrade occurs, we recommend backing up your high-value prompts and outputs.
Save your prompt to a folder
Instead of relying on chat history, use your prompt folders to record:
- The specific instructions you gave
- Any variables or data you provided
- The output you want to reference later
For help setting this up, see How to Create Prompts and How to Manage Prompt Folders.
Summarise and export
Before a session ends, use this meta-prompt to archive your work:
"We have finished our task. Please provide a concise summary of the final results, the key constraints we used, and a starter prompt I can use to resume this project in a new session later."
Manual backups
- Print to PDF
- Markdown Copy
Use your browser's Print (Ctrl/Cmd + P) function and select "Save as PDF" to keep a visual record of the entire thread.
Most code blocks and long-form responses have a Copy button. Paste these into your local editor to keep the formatting intact.
What to Expect After the Upgrade
Once the new model is live:
- Faster response times — Optimised processing power
- Better instruction following — More precise, steerable conversations
- Increased context window — The ability to retain more information within a single session
Run your most important prompts through the new version immediately after an upgrade. You may find that the new model requires fewer instructions to achieve the same or better result.