The lightweight, open-source editor MerMark has just received a major update with version 0.2.0, introducing a built-in AI assistant powered by your existing Claude or Codex CLI subscriptions. Designed for users who prefer simplicity over elaborate workflows, MerMark remains a no-frills editor that combines Markdown editing with inline Mermaid diagram rendering—now with added AI capabilities and enhanced conflict resolution.
A Focused Editor for Markdown and Mermaid Diagrams
MerMark positions itself as a streamlined alternative to feature-heavy note-taking apps. Unlike Obsidian, which thrives on graph views, plugins, and complex automation, MerMark offers a straightforward experience: open a .md file, edit with real-time preview, and render Mermaid diagrams without configuration. Its closest counterparts are editors like Typora, but with a side panel for AI assistance and local-first operation.
The editor continues to excel in baseline features carried over from v0.1.x, including:
- WYSIWYG Markdown editing with toggleable raw syntax view
- Split-screen editing and draggable dividers for multitasking
- Syntax highlighting for over 50 programming languages in code blocks
- Support for tables, task lists, blockquotes, footnotes, and page breaks
- Auto-save, character/word counts, and undo/redo functionality
Mermaid diagram support remains a core strength, with 20+ diagram types rendered inline—from flowcharts and sequence diagrams to Git graphs and mind maps. Users can adjust zoom levels up to 400%, export crisp vector PDFs, and manage diagram scaling to prevent overflow. A new default 25% scale for new diagrams ensures better readability out of the box.
Conflict Detection and Atomic Saves Prevent Silent Data Loss
A standout improvement in v0.2.0 is the introduction of conflict detection and atomic save mechanisms. When external changes—whether from a Git pull, a sync client, or an AI-assisted rewrite—affect a file you’re editing, MerMark freezes and presents a side-by-side diff. You can then decide whether to keep your unsaved changes, overwrite with the new version, or merge both.
This safeguard applies universally, whether the external change originates from your own AI panel or a teammate’s update. The editor also employs atomic writes, using temporary files and verified renaming to prevent mid-save corruption. Users can manually reload files (Ctrl+R) or drag .md files directly into the editor for instant access.
Local AI Integration Without Extra Subscriptions
The headline feature of v0.2.0 is the local AI assistant, which communicates with your existing claude or codex CLI tools—no proprietary API keys or additional billing required. If you’re already subscribed to Claude Code Pro, OpenAI Codex, or ChatGPT Plus, MerMark leverages those credentials seamlessly.
The AI panel can perform a variety of tasks directly in your documents:
- Rewrite sections for tone or clarity
- Extract action items or key points
- Translate notes or content
It operates within a per-document access map, allowing you to grant read/write permissions to specific folders. The assistant can generate summaries, split long documents, or build tables of contents—all while writing changes directly to disk. Optional toggles let you enable shell commands and web searches, though these are disabled by default.
Before every AI write, MerMark creates an auto-snapshot, enabling one-click revert if the output doesn’t meet expectations. A unique advantage is its handling of concurrent edits: if you type while the AI is writing, the conflict dialog activates, letting you choose between your changes, the AI’s edits, or a merged version. This eliminates the frustration of AI overwrites interrupting your workflow.
Users can highlight text and pin it to the AI panel, which then focuses on the selected content rather than the entire document. Screenshots can be pasted (Ctrl+V) or dragged into the chat, and both Claude and Codex support image inputs. Attachments appear as thumbnails in the chat history for easy reference.
Each conversation thread is tied to a specific document and remembers the CLI model and effort level used. If you reopen an old chat, the panel restores the previous settings. Tool calls—such as file reads, writes, shell commands, or web fetches—are logged inline as expandable chips, providing full transparency into the AI’s operations.
Language Support and Performance Enhancements
Version 0.2.0 introduces Chinese as a second bilingual UI option, expanding beyond the original English and Polish support. The editor remains lightweight at approximately 15 MB, built on Tauri rather than Electron, ensuring fast performance without sacrificing features.
Additional improvements include:
- Separate themes for code blocks, allowing syntax highlighting to remain legible in light mode
- Independent zoom controls for the editor and Mermaid previews
- Live token counters for OpenAI, Anthropic, and Google Gemini models
- Keyboard shortcuts that auto-render Mac glyphs on macOS
Looking Ahead: A Tool for Local-First, AI-Assisted Editing
MerMark’s v0.2.0 update strikes a balance between simplicity and power, catering to users who want AI assistance without cloud dependencies or additional costs. By integrating with existing CLI tools and prioritizing local file safety, it carves out a niche for developers, writers, and teams who value transparency and control.
As AI tools become more embedded in workflows, editors like MerMark highlight the potential of local-first, subscription-agnostic solutions. Whether you’re drafting technical documentation, managing meeting notes, or visualizing system architectures, the latest version offers a compelling blend of utility and autonomy.
AI summary
MerMark Düzenleyici’nin yeni versiyonu, yerel AI entegrasyonu, gelişmiş Mermaid diyagramları ve çoklu dil desteğiyle kullanıcı dostu bir deneyim sunuyor. Tamamen yerel ve hafif yapısıyla dikkat çekiyor.
