Visible messages and decisions
Requests, responses, lasting decisions, and next steps stay available with their source.
Verifiable project context
Switch agents. Pick up right where you left off.
Tramya ties every session to the repository, branch, and worktree it belongs to.
What context needs to hold.
A reliable handoff depends on concrete data, not a generic summary.
Work that was actually observed
Commands, changed files, tests run, and visible errors let the next agent verify the state of the work.
Git environment
The repository, branch, and worktree prevent resuming a session in the wrong folder or on a different base.
From lookup to handoff.
Tramya turns the history you already have into a resume point another agent can act on.
Index compatible sources
The local companion detects supported histories without asking for the password to your AI accounts.
Classify every session
Local paths and Git information are used to attach each session to the right project.
Prepare a compact context
The handoff brings together the goal, the observable state, the decisions, and the remaining actions.
Continue in the agent you choose
The new agent loads this context, checks the worktree, and resumes without inventing the missing work.
Questions about AI context
What is an AI context manager?
It is a tool that links a project's sessions, decisions, memory files, and Git information so you can find them again and pass them to another agent.
Does Tramya replace Claude Code or Codex?
No. Tramya keeps and organizes the context produced by compatible agents. The work still runs in Claude Code, Codex, Cursor, or another agent.
Are hidden internal reasoning traces recorded?
No. Tramya keeps only the visible messages, tool outputs, memory files, and traces the source application actually exposes.