Context Connectors open-sources agent indexing layer

This is the kind of release that matters more to builders than to hype cycles: it takes the messy plumbing behind agent context retrieval and turns it into a composable developer primitive.

• –The multi-source story is the headline feature: GitHub, GitLab, Bitbucket, and website indexing in one library is a practical foundation for real-world agent systems
• –MCP support makes the project immediately useful in existing agent stacks, including Claude Desktop, Cursor, and other MCP-compatible clients
• –Local filesystem and S3 storage options push it beyond a toy demo and into something teams can wire into production retrieval pipelines
• –Incremental updates, ignore-file support, and GitHub Actions/webhook recipes show Augment is aiming at continuous indexing, not one-off ingestion
• –The experimental API warning and static-HTML limitation matter: this is powerful infrastructure, but still early and not yet a universal crawler for every JS-heavy knowledge source