Get AI summaries of any video or article — Sign up free
9 New Obsidian Plugins You Should be Using thumbnail

9 New Obsidian Plugins You Should be Using

Prakash Joshi Pax·
6 min read

Based on Prakash Joshi Pax's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.

TL;DR

Harper provides private, fast grammar checking with red underlines and hover-based correction suggestions, but it may be less robust than Grammarly or LanguageTool.

Briefing

A new batch of Obsidian plugins is pushing the app beyond note-taking—adding grammar checking, YouTube-to-notes workflows, local text-to-speech, AI pattern execution, and even startup-time optimization. The most consequential theme across the lineup is practical integration: tools that pull in external content (YouTube, Todoist, Morgan calendar), generate or transform text (AI patterns, text-to-speech), and reduce friction inside Obsidian (lazy loading, random discovery).

Harper is positioned as a free, open-source grammar checker that runs privately and quickly. After configuring grammar rules, it flags errors with red underlines similar to Obsidian’s built-in spelling checker. The difference is interaction: hovering over an error surfaces suggested corrections. While it’s not as robust as Grammarly or LanguageTool—at least based on the presenter’s experience—it remains attractive for users who want an open-source option.

Gulo takes a different approach by turning YouTube engagement into an Obsidian workflow. It fetches all liked YouTube videos, lets users sort them by fields like title, posted date, like-to-view ratio, view count, and like order (ascending or descending), and supports search. The standout feature is portability into daily notes: videos can be added directly to a Daily Note or dragged into other notes. Setup is more involved than typical plugins because it requires Google Cloud Console credentials and the YouTube Data API, including client ID/client secret and a Google login tied to a test user.

Todoist Project Sync is aimed at users who already manage work in Todoist. It creates a note file per Todoist project and keeps those notes aligned with Todoist changes—archiving notes when a project is deleted and restoring them when a project returns. Crucially, it’s one-way: synchronization flows from Todoist to Obsidian, not back.

Listen Up brings local text-to-speech to Obsidian using Piper (open source). The plugin requires selecting a Piper executable, downloading a language model, and creating a configuration file from provided settings. Once configured, text can be converted to speech via the command palette, with troubleshooting routed to the plugin’s Discord if conversion fails.

Stif adds a lightweight “serendipity” mechanic by showing a random note from a chosen source directory every time the vault opens. Users can also exclude directories (like templates or people folders) to control what gets surfaced.

Lazy Plugin Loader targets performance. Instead of loading every plugin at startup, it delays plugin initialization—configurable separately for desktop and mobile. The result is a noticeably faster vault launch, with plugins appearing after the delay.

Quick Cards displays notes in a card-style browser by folder, letting users open and create notes from within that interface.

Mesi is the most complex integration, built around AI “patterns” and running them through multiple providers. It replaces the multi-step setup of the Unofficial fabric integration by letting users configure API keys and models (including options like Llama 3 variants and local endpoints for AMA). It supports downloading Fabric patterns, selecting input context (active note, clipboard, or tab), and running workflows that can chain patterns—producing both an intermediate essay and a final “extract wisdom” output.

Finally, Morgan Task is an official Morgan team integration for time-blocking and planning, but access is limited to premium users. For those eligible, it connects Morgan’s calendar tasks directly into Obsidian’s planning flow.

Taken together, these plugins show Obsidian becoming a hub for external media, local automation, AI-assisted writing, and performance tuning—without requiring users to leave their vault for every task.

Cornell Notes

The plugin lineup expands Obsidian into a more connected workspace: grammar checking, YouTube archiving, local text-to-speech, random note discovery, and AI-driven note generation. Gulo turns liked YouTube videos into sortable lists that can be added to Daily Notes, but it requires Google Cloud Console setup and YouTube Data API credentials. Listen Up uses Piper to convert text to speech locally, requiring a Piper executable, a downloaded language model, and a configuration file. Lazy Plugin Loader improves responsiveness by delaying plugin startup, while Stif and Quick Cards add discovery and folder-based card browsing. Mesi brings AI pattern execution into Obsidian, chaining Fabric-style patterns to generate outputs like essays plus extracted “wisdom.”

How does Harper deliver grammar suggestions inside Obsidian, and what tradeoff does it have versus tools like Grammarly or LanguageTool?

Harper is free and open source, designed to run privately and quickly. After configuring grammar rules, it marks issues with red underlines similar to Obsidian’s default spelling checker. Hovering over an underlined error shows suggested corrections. The tradeoff mentioned is that it isn’t as robust as Grammarly or LanguageTool, likely because those tools rely on larger databases and resources.

What makes Gulo useful for turning YouTube activity into a note workflow, and why is setup more complex than typical plugins?

Gulo fetches all YouTube videos a user has liked and provides sorting by fields such as title, posted date, like-to-view ratio, view count, and like order (with ascending/descending options). Users can search within the list and then add videos to a Daily Note or drag them into other notes. Setup is more complex because it requires Google Cloud Console credentials and the YouTube Data API, including creating a project, enabling the API, generating client ID/client secret, and configuring authorized JavaScript origins and redirect URLs.

What does Todoist Project Sync synchronize, and what limitation prevents it from being a two-way integration?

Todoist Project Sync creates a note file for each Todoist project and keeps those notes aligned with Todoist changes. When a Todoist project is deleted, the corresponding Obsidian node is archived; when restored, the node is restored from the archive. If a project is moved or renamed, the corresponding note and subnotes are moved/renamed too. The limitation is one-way sync: it syncs from Todoist to Obsidian only, not from Obsidian back to Todoist.

What are the practical steps and dependencies behind Listen Up’s local text-to-speech?

Listen Up relies on Piper, an open-source text-to-speech system. In plugin options, users must provide the Piper executable file, a custom model file, and a model configuration file. Users download the Piper executable for their OS (Mac/Linux/Windows) and select a language model (e.g., English). The configuration file requires copying provided configuration code into a new file and pointing the plugin to it. Conversion is triggered via the command palette, and if it fails, the plugin’s Discord is suggested for troubleshooting.

How do Lazy Plugin Loader and Stif change day-to-day vault behavior?

Lazy Plugin Loader reduces startup overhead by loading plugins after a delay instead of immediately. It supports separate desktop and mobile configurations and can set a short delay (e.g., 10 seconds) for new plugins and a longer delay (e.g., 20 seconds) for others, improving vault launch time. Stif changes vault opening behavior by showing a random note popup every time the vault opens, chosen from a configurable source directory while respecting excluded directories like templates or people folders.

What does Mesi add beyond the Unofficial fabric integration, and how do chained patterns affect outputs?

Mesi is designed to streamline AI pattern workflows that otherwise require multiple steps with the Unofficial fabric integration, including terminal setup and Fabric configuration. It lets users configure provider/API keys and models (including Llama 3 variants and local endpoints for AMA via a localhost URL). It can download existing Fabric patterns, select input context (active note, clipboard, or tab), and run patterns from a patterns folder. When patterns are chained (stit), the AI can generate an intermediate result (like an essay) and then produce a final output (like “extract wisdom”) in a new note—showing both stages rather than only the final extracted result.

Review Questions

  1. Which plugins require external credentials or cloud configuration, and what specific services do they depend on?
  2. Compare how Lazy Plugin Loader and Stif each affect vault startup—one targets performance, the other targets discovery.
  3. Describe how Mesi’s pattern chaining changes what ends up in the final note compared with a single-pattern run.

Key Points

  1. 1

    Harper provides private, fast grammar checking with red underlines and hover-based correction suggestions, but it may be less robust than Grammarly or LanguageTool.

  2. 2

    Gulo turns liked YouTube videos into sortable lists and lets users add them to Daily Notes or other notes, but it requires YouTube Data API credentials via Google Cloud Console.

  3. 3

    Todoist Project Sync maintains one-way synchronization from Todoist to Obsidian by creating per-project notes, archiving on deletion, and restoring on reactivation.

  4. 4

    Listen Up enables local text-to-speech using Piper, requiring a Piper executable, a downloaded language model, and a properly created configuration file.

  5. 5

    Stif improves note discovery by showing a random note popup on vault open, with configurable source and excluded directories.

  6. 6

    Lazy Plugin Loader speeds up Obsidian startup by delaying plugin initialization, with separate desktop/mobile delay settings.

  7. 7

    Mesi brings AI “pattern” execution into Obsidian, supporting multiple providers/models and chained workflows that can generate both intermediate and final outputs.

Highlights

Gulo can pull every liked YouTube video into Obsidian and drop selected items into Daily Notes—but only after YouTube Data API credentials are configured.
Listen Up performs text-to-speech locally via Piper, avoiding server calls but requiring careful setup of executable, model, and configuration files.
Lazy Plugin Loader can make vault startup feel “instant” by delaying plugin loading, then adding plugins after a timed delay.
Mesi supports chained AI patterns, producing multi-step outputs (an essay followed by extracted “wisdom”) inside new Obsidian notes.
Stif surfaces a random note every vault open, using directory include/exclude controls to steer what gets discovered.

Topics

Mentioned

Get summaries like this for any content

Videos, articles, research papers — all summarized by AI and searchable in one place.

Start building your library