Pieces for Developers Documentation

The Pieces CLI is a command-line tool designed for interacting with Pieces OS. It provides programmatic access to core Pieces functionalities, including:

• Code Snippet Management: Manage and interact with your saved code snippets directly from the terminal.
• Copilot Interaction: Engage with Pieces Copilot features via the command line.
• Feature Integration: Access and integrate various other Pieces OS features into your development workflow.

You can save various developer materials into Pieces for Developers using universal methods that work across most interfaces, or specific methods tailored to the platform you are using (Desktop, IDE plugins, or Web Extension).

Supported Materials

• Images: .png, .jpeg, .jpg, .gif
• Raw Text or Code
• Whole Files: Importing a directory allows Pieces to use Snippet Discovery to find individual snippets within that directory.
• GitHub Gists: Imported via Snippet Discovery.

Pieces OS is a background service that acts as the core engine for the Pieces ecosystem. It enables local operations and handles communication between various integrations.

Key functions of Pieces OS include:

• Enabling Pieces to work locally.
• Managing communications with installed plugins.
• Allowing Pieces to save snippets.
• Generating responses to user questions.
• Enabling cloud-enabled features like link sharing.

Pieces OS is the essential background service that enables Pieces to function locally. It handles all communications with integrations (like the Azure Data Studio extension).

Key functions of Pieces OS:

• Enables saving of code snippets.
• Facilitates interaction with IDE plugins.
• Enables cloud synchronization for data access across devices.
• Powers cloud-enabled features such as link sharing.

Note: Without Pieces OS running, the Azure Data Studio extension will not be able to save snippets or interact with your workflow.

Pieces OS is a background service that runs locally on your machine to facilitate communication between products in the Pieces for Developers Suite. It acts as the engine for the ecosystem, providing:

• Local Execution: Enables products to operate 100% locally for privacy and speed.
• Machine Learning Management: Manages local and cloud-based ML models used by Pieces Copilot.
• Data & Search: Handles real-time search, suggestions, and data processing workflows.
• Cloud Connectivity: While primarily local, it can connect to the cloud for backup, snippet sharing, and specific cloud-only ML models.

When installing Pieces for Developers, you are installing two distinct but interconnected components:

1. Pieces OS: A required background service that runs offline and on-device. It acts as the centralized storage agent that enables communication between different products in the Pieces suite. It manages local, private, and secure Machine Learning models, real-time search, and suggestions.
2. Pieces Desktop App: The flagship visual GUI application. It provides the user interface for navigating snippets (via Gallery or List views), transforming code between languages, sorting snippets by AI-suggested usefulness, and interacting with the Pieces Copilot.

Pieces OS is required for the suite to function, providing the local infrastructure for LLMs and data synchronization.

The Long-Term Memory feature is designed to be private by performing all data aggregation and LTME processing locally on your device.

Privacy Recommendation: To ensure that no data leaves your device during the LLM interaction phase, it is strongly recommended to use a Local LLM (such as Mistral, Phi-2, or Llama2) with the Pieces Copilot. This ensures that both the context retrieval and the language model processing happen entirely offline/locally.

In collaborative workflows where project context (decisions, architectures, implementations, and configs) is scattered across channels like Slack, Teams, or Discord, Pieces for Developers automates context preservation.

When you save code snippets, the platform uses a Context Awareness Engine to automatically extract and associate over 15 metadata heuristics with that snippet. This ensures that the 'who, what, where, and why' behind a piece of code is preserved alongside the code itself. Additionally, the Pieces for Developers Desktop App allows you to toggle realtime and scope-relevant suggestions to help you find materials to save, reference, reuse, or share.

When you save a snippet, Pieces automatically processes it to add the following context (enrichments):

• Programming Language: Automatically detected.
• Title: A generated title for the snippet.
• Tags: A set of generated tags for categorization.
• Description: A smart description explaining what the code does.
• Search Suggestions: Suggestions for searches to help you learn more about the concepts in the snippet.
• People/Authors: Information about people related to the snippet, extracted from Git history if the file is within a Git repository.