Run History & Search

Mochi learns what you actually run. Every run is recorded, search ranks by it, favorites pin it, and stats show it back to you.

Runs

Every run is kept, whether you started it in the desktop app, typed it in a terminal, or an agent ran it over MCP. The Runs page in the desktop app lists them all, live and past, with their output.

The Origin column names the client that started a run (Claude Code, Cursor, Mochi's own chat), and the Agent runs tab lists only those. Filter History by Source (CLI, Desktop or Agent (MCP)), or by client and session.

When a run fails, open it and ask Mochi to diagnose it.

flow conceptHistory & logs

Every run is a flow execution, and flow keeps the logs. mochi logs works exactly like flow logs.

Search

mochi search            # your favorites and most recent executables
mochi search deploy     # ranked matches across every workspace

Results are ranked by how well the name matches, whether it's a favorite, and how often and how recently it has run. The desktop's ⌘K palette uses the same search.

Flag	Effect
-n, --limit	Maximum results (default 20)
-o json	Machine-readable output

Favorites

Favorites sort first and show up on the desktop home screen. They're shared between the CLI and the desktop app.

mochi favorite add --workspace my-project
mochi favorite add --executable "build my-project/app:image"
mochi favorite list
mochi favorite remove --executable "build my-project/app:image"

To reorder favorites, drag them in the desktop app.

Stats

mochi stats               # every workspace
mochi stats my-project    # one workspace

Shows total runs, success rates, and the busiest executables. -n sets how many top executables to list, and -o json gives machine-readable output. A run you stopped (Ctrl-C, or Stop in the desktop app) counts as canceled rather than failed, and stays out of success rates and average durations.

The desktop's Insights page builds on the same numbers, and mochi ai insights turns them into recommendations.