Exploring The Pieces for Developers AI App - My Initial Thoughts

These days we are so fortunate to live in a time where we have access to dozens of applications which can dramatically boost our productivity. I use quite a few popular productivity tools because I follow the Getting Things Done (GTD) workflow which is essentially a personal productivity philosophy that reinvents how you approach your life and work.

Today I want to introduce you to one of the hottest new AI-enabled productivity tools around called Pieces. It has completely transformed my productivity workflow and I rate it as highly as other productivity tools like Notion and Obsidian which I am already using daily. I am getting so much usage out of it that I added it to the dock on my Mac right next to Visual Studio code 😄

I suppose you are wondering what is Pieces exactly.

What is Pieces?

The Pieces team say that:

"Pieces is your AI-enabled productivity tool designed to supercharge developer efficiency. Unify your entire toolchain with an on-device copilot that helps you capture, enrich, and reuse useful materials, streamline collaboration, and solve complex problems through a contextual understanding of your workflow."

Basically, this means that you can use the Pieces desktop application to save code snippets and use its AI features for various tasks which I'm sure you are already familiar with if you use ChatGPT a lot. You get access to Large Language Models (LLMs) like OpenAI, Gemini and others both in the cloud and locally on your machine.

You can download the Pieces application from their website and see just how well it can improve your workflow.

I was quite fortunate to discover this application. For a while, I had been seeing mentions of it on X formerly Twitter but I thought it was related to GitHub Copilot when it is so much more. Arindam Majumder reached out to me on X formerly Twitter and this led to some coffee chats with Rosie Levy and Tsavo van den Berg who work for Pieces. I played around with the app and my first impression was that it felt like a product that could be built into a future version of an operating system one day that's how useful it is.

Being able to switch between multiple different applications, plus using context and having Pieces know what you're working on while offering AI assistance and help can feel like real magic. It's almost like having your own personal Cortana, Jarvis, Lucy, Siri with superpowers or whatever future AI avatar you can imagine from a movie or real life. This is only the beginning and I fully expect it to get better over time because AI tools like ChatGPT have only been around since 2022.

Anyway, today I will go through a high-level overview of what my first impressions were when I started to use the desktop application. I will go through the 7 main application features one by one which are:

• Saved Materials
• Captured Context
• Copilot Chats
• Workflow Activity
• Global Search
• Snippet Discovery
• Updates & Upcoming

Alright, let's start with Saved Materials.

Saved Materials

You can save code snippets by adding them to the application or by using one of the browser, IDE or other extensions. It's possible to copy, edit and manage the code snippets in different ways. It's even possible to use the inbuilt copilot to start a conversation about the code snippet which is very cool and a great feature.

Captured Context

The captured context section seems to capture data on relevant information. This can include websites you have visited, people you interacted with, different files and also relevant tags. I guess you could think of it like a smart AI book marker that compiles information that you might find useful based on what you are doing.

Copilot Chats

This part of the application is the most familiar to most people. If you have used an AI tool like ChatGPT, Copilot, Gemini, Claude or any of the others then you already know what to expect. You use the prompt to ask the AI questions and then it responds. The Pieces desktop app is unique though because you can choose to use different LLMs. As you can see here I am using GPT 4 however the latest GPT-4o is also available to use.

Here you can see the Pieces Copilot LLM selection screen.

Workflow Activity

The workflow activity does exactly as you would expect it to do. It tracks everything that you have been doing and shows you a timeline of the events. You can go through the files and locate the activity within the Pieces application.

Global Search

When using the global search you can search for anything inside of the application that you have worked on. This includes chats, snippets and many other things.

Snippet Discovery

In this section of the Pieces application, it becomes possible to add entire codebases to our snippet database. All you have to do is select a codebase and then Pieces will import the relevant files. You then get the option to select which files you want to import.

Updates & Upcoming

From this section, we can see the upcoming roadmap for the Pieces application. There are links to the discussions on GitHub so you can take part and see what the progress is with the latest feature requests. if the feature is already available then you can just follow the link to the feature and it will take you there inside of the Pieces application.

Final Thoughts

We have just barely scratched the surface here the Pieces application is capable of doing so much and it is only going to get better. I am enjoying my experience so far and this has become one of my must-have tools as a developer. I can see the Pieces application getting mentioned in the same conversations as Visual Studio Code, Notion and Obsidian.

As of writing, there are plugins for Visual Studio Code and Obsidian and in the future, there could be one for Notion as well. I posted some videos on X formerly Twitter that show the Visual Studio Code plugin in action as well as the Live Context feature. The Live Context feature essentially gives you real-time contextual information throughout all applications on your operating system. So it can see what you are writing in documents, browsing on the web, the code you are writing, emails and much more.

It's worth mentioning that the application is privacy-orientated so all of your data and private information should be safe. You can turn off the Live Context mode if you want to.