Overview
Semlith Cloud keeps one index of the code, docs and tickets you connect to it. It reads from GitHub, GitLab, Confluence and Jira, and from folders you push from a laptop or upload. This connector lets Claude search that index and read what it finds, so every answer points at the file and line it came from. Claude can search by meaning or for an exact string, read a whole file or one function, list a folder, find where something is defined, see what calls it and what it calls, and check every place a change would reach. Answers come back as short ranked spans, not whole files. The connection is read only. Claude can't index, add or remove anything through it. No model is trained on your content. You need a Semlith Cloud account with at least one source connected. When you connect, you sign in with GitHub or GitLab, or your team's SSO, and pick the workspace Claude should search. You can disconnect it at any time from the Agents page in Semlith Cloud. What Semlith keeps, and for how long, is on https://cloud.semlith.com/data.
