Sets

Learn about Sets — curated collections of knowledge versions in Bindly. Token budget retrieval for AI context management.

A Set is a curated collection of specific Binding Versions — like a playlist for knowledge.

Why Sets?

Sets let you:

  • Group related knowledge for a specific context or project
  • Pin specific Versions (not just "latest") for stability
  • Retrieve all content at once with a token budget via mcp_get_set_context

How Sets Work

A Set contains references to specific Versions, not Bindings. This means:

  • When a Binding is updated, the Set still points to the Version you added
  • You control exactly which Version of each Binding is in the Set
  • To update, remove the old Version and add the new one
Set: "Deployment Reference"

├── Position 1: Kubernetes Guide v3

├── Position 2: Docker Best Practices v1

└── Position 3: CI/CD Pipeline Notes v2

Token Budget Retrieval

The most powerful Set feature is mcp_get_set_context:

mcp_get_set_context({

setId: "set-id",

maxTokens: 8000,

tier: "1.5"

})

This assembles all Versions in the Set within your token budget. If the total exceeds the budget, content is truncated and truncated: true is returned.

Creating and Managing Sets

Create a Set

mcp_create_set({

spaceId: "space-id",

name: "My Collection",

description: "Related resources for project X"

})

Add Versions

mcp_add_version_to_set({

setId: "set-id",

versionId: "version-id",

position: 1

})

Remove Versions

mcp_remove_version_from_set({

setId: "set-id",

versionId: "version-id"

})

Set Metadata

Each Set tracks:

  • versionCount: Number of Versions included
  • totalTokenCount: Sum of all Version token counts (for budget planning)