Watching it work

Most of a turn is the agent doing things rather than talking. Semantix spends that time telling you what's happening — in the transcript, in the editor, and in the 3D view of your project.

The indicator#

While the agent works, one line sits above the message box. On the left, the Semantix mark and a word — Warming up, Cogitating, Skimming, Etching, Sleuthing, Churning, Composing and a couple of dozen others — often with a detail trailing it: the file being read, the pattern being searched for, the command being run. On the right, the meters: how many seconds the current tool has been running, and on Claude Code a live token count for the turn.

The word tells you the phase, not the specific action. Each phase — booting, reasoning, reading, writing, searching, listing, running, responding — has its own small pool, and one is drawn at random and re-rolled as it shimmers. Two different files being read look the same apart from the trailing filename. The phase itself is real; it's driven by the tool the agent actually called.

If the model goes quiet composing a long tool call, the detail becomes preparing write file (6s) and the seconds tick up. A frozen number there means the run has died.

The transcript also gets a collapsible thinking row in every turn where the model reasoned — click its header to read. The header word is chosen once per turn and stays put. This is the same on every provider, Claude Code included.

Following the agent#

The eye icon in the strip above the message box opens a three-way picker:

• Visual follow (3D camera) — the default. The camera flies to whatever the agent is touching.
• Editor follow (open agent's file) — a tab opens on the file the agent is working on.
• Follow off.

The choice is remembered across projects.

In editor follow, one tab is reused as the follow channel: a second file closes the first and retargets. If you had already opened that file yourself, your tab wins and is simply focused.

In the 3D view, the camera moves to whatever the agent touches, and a card attaches to the node showing the lines going past — labelled writing or editing while it happens, wrote or edited once it's done. A file that doesn't exist in the scene yet appears as a ghost node that becomes real when the write lands; a file already on screen flashes. When the agent searches, each hit is struck by a bolt from its folder and lights up as it lands; when it lists a folder, the folder and everything found in it light up together, and the rest of the scene dims.

Up to three cards at once. A write card fades a second or two after it lands; a read card lingers around five seconds.

Two more things about the camera. Touching it takes it back — orbit or drag during a turn and the agent stops steering it for that step. And in visual mode the camera returns to an overview of the project at the start and end of each turn, which will undo a folder you had flown into. Folders the agent opened along the way stay open; nothing folds by itself.

Notices in the transcript#

Some things are reported as a centred line rather than as chat: Stopped, Response truncated when the model hit its output limit, Response filtered when a safety filter cut in, Agent error with the message, and Model when you switch the Claude model mid-conversation. Memory reports the same way — the walker refusing for lack of a seat, or a hand-off landing.

They're informational and they persist with the conversation.

Stopping#

The send button becomes a stop button while the agent works. Press it, or press Escape in the message box. You get a Stopped line, the indicator clears, and anything you had queued is dropped.

Two honest caveats.

Denying a permission also ends the turn, without a Stopped line — the denial row says it. Closing a chat tab stops its run silently, since there's no longer anywhere to show the notice.