## n8nChat Changelog

Track our latest updates and improvements to n8nChat.

### v0.4.2

**January 8th, 2026**

• **Added new models:**
  
  -- GPT-5.2  
  -- GPT-5.2-Pro  
  -- Gemini-3  
  -- Gemini-3-Flash

### v0.4.1

**September 30th, 2025**

• Reduced token usage by ~30%  
• Increased speed of workflow generation  
• Higher output quality when working with complex nodes, like the AI Agent node  
• Fixed a bug that caused messages to fade away while the chat was still responding  
• Fixed a bug which caused the stop button to not work  
• Fixed typos in the settings FAQ  
• Fixed issues with the conversation getting stuck without properly displaying errors in chat

### v0.4

**August 7th, 2025**

• **New OpenAI Models added: gpt-5, gpt-5-mini, and gpt-5-nano**  
• Gemini models switched from preview versions to main versions  
• 'Support' tab added to n8nChat settings menu, containing frequently asked questions and support links

### v0.3.3

**July 7th, 2025**

• Fixed a bug that would log users out of their n8nChat account after 1 hour

### v0.3.2

**June 23rd, 2025**

• **Gemini 2.5 Flash Lite has been added**  
• **All Gemini models now support Google Search.** You'll see the toggle for this setting in the model section (on by default) and it allows n8nChat to search the web whenever it determines it needs up-to-date information to complete your request  
• All Gemini models will use less tokens, especially when working with longer conversations  
• A new setting has been added which allows you to set the delay before new messages fade away. You can set this to 0 to turn off the effect completely  
• The default settings for temperature, top p, and max tokens have been adjusted  
• Fixed a bug that caused a disconnected port error to appear on the first message in n8n Chat with a fresh workflow page load  
• Fixed a bug that prevented some Gemini error messages from appearing in chat (usually rate limit errors) and instead returned blank  
• Fixed a bug that made it difficult to adjust the API sliders  
• Fixed a bug that prevented you from using the backspace key when an image was attached to the chat  
• Various tooltips and error messages have been updated to be more helpful  
• Missing translations for some languages have been added

### v0.3.1

**June 13th, 2025**

• **New feature added to early access for Premium n8nChat subscribers: Custom system prompt instructions. This feature will be coming to all users in a future update, but for now is exclusive to those who have supported development with a subscription - thank you!**  
• OpenAI o3 added  
• Image support added for o4-mini  
• Fixed bug with n8nChat UI not displaying in some cases  
• Fixed OpenAI error messages sometimes not displaying in chat (rate limit errors, etc)  
• Added missing text translations for other languages

### v0.3

**May 31st, 2025**

• **Google Gemini support added**  
  -- Gemini 2.5 Pro  
  -- Gemini 2.5 Flash  
  -- Gemini 2.0 Flash  
• Google is now the default Provider and Gemini 2.5 Flash the default model  
• Image support added. Attach a screenshot of any workflow and simply ask n8nChat to recreate it for you!  
• n8nChat now knows what version of n8n you're using  
• API error messages now correctly display in-chat  
• Fixed bug causing blank responses  
• Fixed bug causing permanent "Thinking" status  
• Fixed bug that would log you out of your n8nChat account after an hour. You should now remain logged in indefinitely

### v0.2.7

**May 10th, 2025**

The Firefox version is now updated to this version  
• Added Vietnamese localization  
• Fixed a bug in node creation that caused old, outdated versions of nodes to be created instead of the latest version  
• Fixed blank messages caused by API errors. Errors will now always show in the chat  
• Fixed login issues (Firefox)  
• Fixed crashing on self hosted installs (Firefox)

### v0.2.6

**May 7th, 2025**

Thank you for all of your feedback via the Discord and extension reviews! They've helped greatly in making the extension better for everybody.  
• Agentic self-correcting capability added. Now when n8nChat creates nodes, it'll automatically check them for any issues like missing or invalid node parameters and attempt to correct them itself. If it requires manual input, it'll create a sticky note step-by-step instructions  
• Default AI model switched from o4-mini to gpt-4.1. Many users were finding the reasoning models required account validation from OpenAI before they could be used. This switch is an equally impressive model without requiring account validation, making it easier to get started with n8nChat  
• Added a deactivate button to remove n8nChat from the page  
• Fixed duplicated messages after switching models  
• Fixed chat input bar spacing  
• Removed gpt-4.1-nano (this model just isn't good enough for these kind of tasks)  
• Change log link added to extension settings menu

### v0.2.5

**May 1st, 2025**

• Improved markdown formatting so AI messages are easier to read  
• Fixed text being duplicated when using gpt-4.1-mini/nano models  
• Fixed the quality of gpt-4.1-nano output so it doesn't get stuck  
• Improved API error handling, so if the API responds back with an error it'll show in the chat UI as a message instead of just going blank  
• Added setup tutorial video link to the API settings  
• Added localization support, languages now supported: Portuguese, Spanish, French, German, Polish, Italian, Korean, Chinese, Russian

### v0.2.4

**April 26th, 2025**

• Fixed extension crashing on self-hosted n8n installs  
• Fixed repeating text in some AI responses  
• Fixed API errors  
• Improved token efficiency and reduced token usage

### v0.2.3

**April 22nd, 2025**

• New models added: gpt-4.1-nano, gpt-4.1-mini, o4-mini, o3  
• Bug fixes

### v0.2

**April 18th, 2025**

• Bug fixes

### v0.1

**April 14th, 2025**

• Initial release