Connection overview

Connections bring source material into ctx| so agents can use it as organization-scoped context. The important thing to understand is that ctx| is Git-first: GitHub gives ctx| access to code, and Git repositories also provide a transparent target for content synced from other systems.

That means your team keeps control. You choose what ctx| can read, where synced content lands, and what becomes part of the knowledge system agents can query.

GitHub is the foundation

Most teams start by connecting GitHub. This lets ctx| read selected repositories, build code search, and ingest repository structure into your organization's knowledge system.

GitHub is critical because it supports two jobs:

• It gives ctx| direct access to source code for single-repo and cross-repo code search.
• It provides the target repositories used by other connections, such as Confluence, when external content needs to become repo-backed context.

Other connections flow through Git

For docs and tools outside GitHub, ctx| keeps synced content visible to your team. The connection scopes the source first, then syncs selected content into a GitHub repository you control.

For example, the Confluence connector lets you choose spaces and pages, copy that content into a selected repository in your GitHub organization, and then ingest from that repository into ctx|. When upstream Confluence content changes, the connector can update the Git-backed source so ctx| can ingest the new state.

This gives you:

• A visible Git history for synced context.
• A clear repo owner and review surface.
• One ingestion path for code, docs, and synced source material.
• Better control over what agents can rely on.

Agents use ctx| through MCP

MCP is agent access, not another source connection. Agents connect to ctx| once, then ask for context from the sources your organization has connected, indexed, and ingested.

For setup details, see ctx| MCP.