Skip to content
Wire Docs

Container Best Practices

This guide covers best practices for organizing context containers.

Planning Your Containers

Section titled “Planning Your Containers”

Before creating containers, consider:

  1. Scope - What context belongs together?
  2. Access - Who needs access to this context?

Container Names

Section titled “Container Names”

When you create a container, Wire assigns a random placeholder name like “Prancing Zebra”. You can change the name and description anytime in Container Settings.

Naming Best Practices

Section titled “Naming Best Practices”

Good container names are concise (1-3 words), describe the content domain, and slugify well (no special characters).

Examples: “Product Docs”, “Customer Support”, “Sales Playbook”

The container name is used for MCP client naming (e.g., “wire-product-docs”) but does NOT affect the MCP URL (which uses the container ID).

Organization Patterns

Section titled “Organization Patterns”

By Project

Section titled “By Project”

One container per project or product:

my-org/
├── web-app-docs
├── mobile-app-docs
└── api-specs

By Team

Section titled “By Team”

One container per team or department:

my-org/
├── engineering
├── design
└── customer-success

By Content Type

Section titled “By Content Type”

One container per content type:

my-org/
├── documentation
├── research-data
└── meeting-notes

Common Mistakes

Section titled “Common Mistakes”

Too Many Small Containers

Section titled “Too Many Small Containers”

Problem: Creating a container for each document.

Solution: Group related documents in one container. With too many small containers, relationships between data aren’t discovered across containers, so your agent has to build that context at query time.

Too Few Large Containers

Section titled “Too Few Large Containers”

Problem: Putting everything in one container.

Solution: Separate by topic or access needs. Purpose-driven containers offer better context and more relevant search results, making it easier for your agent to execute and saving tokens.

Next Steps

Section titled “Next Steps”