EmEditor

[Camel1965]

EmEditor Professional 25.3.0​

• Generative AI Enhancements (ChatAI plugin required):
• One of the biggest additions in v25.3 is enhanced generative AI support, developed by Makoto Emura. This release adds Google as a provider and supports OpenAI’s GPT‑5. EmEditor now also supports Tool Calling, which lets the AI invoke a limited set of built-in EmEditor tools. It’s similar to an AI agent, but restricted to tools that EmEditor exposes—hence the name “Tool Calling.” With this, you can run natural-language commands on the currently open document via the Chat with AI interface. For example, if you open a file with email addresses and type “Extract email addresses from this document” into the Chat with AI bar, EmEditor will extract them and create a new document.
• To use Tool Calling, click the “Tools” button and ensure it’s highlighted in blue. You can also choose which tools are enabled on the Tool Calling page in Chat with AI settings.
• The Chat with AI Settings and the AI Options page in the Customization dialog let you set Max tokens and Temperature (randomness). Previously, the default Max tokens was 4096, which could cause truncated output for reasoning-heavy models like GPT‑5. In this version, you can toggle Max tokens on or off. When it’s off, the provider’s default is used.
• In Chat with AI Settings, you can also adjust Reasoning effort where supported. When Reasoning effort is off, the model’s default is used. Higher effort yields more reliable responses but can be slower. If you want the fastest replies, turn Reasoning effort on and select Minimal.
• Sort by Similarity:
• EmEditor includes many sorting commands, and this version adds the ability to sort by similarity to the text on the current line or in the current cell. Similarity is calculated using the Levenshtein distance (edit distance) introduced in v22.0.0, with an option to ignore case. For example, if you have a CSV roster and search for Michael, placing the cursor on Michael and running Sort by Similarity will sort the list so close matches like Micheal and Mike appear first—making similar names easier to spot.
• Markdown/HTML Toolbar:
• We’ve added HTML features to the Markdown toolbar introduced in v24.4 and combined them into a single Markdown/HTML toolbar. Markdown and HTML modes switch automatically. Opening a Markdown file and choosing Markdown settings or turning on Markdown Design Mode enables Markdown mode; opening an HTML file and choosing HTML settings enables HTML mode. Accordingly, the old HTMLBar plugin has been discontinued. The new HTML mode is more capable than HTMLBar. For example, the drop-down for Paragraph, Heading 1, Heading 2, and more automatically detects and shows the current style, and buttons like Bold, Italic, and Code reflect the current font state. If no text is selected and you click Bold, the word at the cursor becomes bold. Selecting the Bold command (Ctrl+B by default) while selecting text produces **text** in Markdown mode and <strong>text</strong> in HTML mode.
• Performance and More:
• We continued optimizing startup performance in this release. The Extract Frequent Strings command is also significantly faster. In our tests, startup is 1.90× faster than v25.0, and Extract Frequent Strings is 1.42–4.44× faster than v25.2.
• Other improvements include:
• Responding to a customer request, you can now add conditions to regex highlight strings with special keywords like (?#_text_c!=17) for more flexible highlighting.
• Extract Frequent Strings now respects filters and works on filtered content.
• Previously, the cursor position didn’t change after sorting. Now it moves to the corresponding item, making it easier to run the new Sort by Similarity followed by additional sorts within Advanced Sort.
• In addition to the Customize dialog, the Customize Macros dialog and Configuration Properties now show an asterisk (*) next to page titles that differ from default settings.
• We also fixed issues found during the preview build. Starting with this version, the trial period before registration changes from 30 days to 7 days.
• We hope both the Pro and Free editions meet your expectations. If you have questions, feature requests, or ideas, please contact us or post on the forum.

Zaloguj lub Zarejestruj się aby zobaczyć!