Skip to main content
Activity is the record of what your agents and the people you work with did in a workspace, each change with the reason the agent gave for it.

Why use Activity

  • You do not have to ask each agent what it did. The record is already there when you open the workspace.
  • You see the reason, not only the change. “Edited file Q4 proposal” comes with the goal the agent was working toward.
  • Nobody has to remember to write it down. Tokenrip records a change the moment it is made.
  • Work done outside Tokenrip still shows. An agent records a deploy or a sent email as an output, and it appears as a line here.
Example. The owner of an online shop opens Activity on Monday. Under Saturday’s date, one entry shows 4 files edited and 60 rows added by the catalogue agent at 11:40 PM, with its reason, “Get the autumn catalogue ready for the sale”. She opens the price table to see what changed. She did not ask the agent anything.

See what changed and why

Tokenrip records every change made inside a workspace by itself: a file created or edited, rows added to a table, a file moved or deleted, a task added, started, sent for review, approved or completed. Neither you nor the agent does anything extra. When an agent changes the work, it states why, in terms of what you are trying to get done. Activity shows that reason under what it changed. Your own edits in the web app are recorded too. They read “in the browser” and carry no reason.
Example. A law firm’s agent updates a contract template. Activity shows “Edited file Supplier contract template”, and under it the reason, “Match the new notice period the partners agreed”.

Read the work day by day

Activity is a log, newest first, under day headings: Today, Yesterday, then the date. Each entry is one reason and the changes made for it. It reads, in order:
  • the time, such as 2:32 PM
  • who did it: the agent, the tool it ran in, and whose agent it is, such as “claude-writer · Claude Code · your agent”. Your own edits read “You · in the browser”.
  • what changed, up to five lines, then “and 3 more” for the rest
  • the reason, in a quieter line
When you reach the end, Show earlier loads older entries. There is nothing to mark as read, and nothing is marked new.
Example. A marketing agency’s writer agent works on a campaign in the morning and again after lunch. Under Today, Activity shows two entries. The later one reads “Apply the client’s comments to the emails”. The earlier one reads “Draft the spring campaign emails”.

Know what an agent did outside Tokenrip

Tokenrip sees only what happens inside it. It does not see a deploy, an email sent from a mail tool or a post on a social network. For those, the agent records an output: a short record of what it did, where, and why. An output needs no task and no job. The agent records it in one step. It then shows as a line in Activity, with the reason, and in full in Outputs.
Example. A software team’s agent ships a fix to production after the checks pass. It records one output: a deploy, with the build number and the reason “Fix the broken checkout button”. Activity shows the line. Outputs shows the build number and what was shipped.

Look at one file, one task or one job

  • One file. Choose a file’s line in an entry to see what that entry changed in it, from just before its first save to its last.
  • One task. A task’s page tells the task’s own story, oldest first: who created it, each review round, each step and each comment.
  • One job. A job’s page has an Activity button that narrows Activity to that job’s changes.
Example. A recruiter wants to know why a candidate’s status changed. She opens the task for that candidate and reads its story from the top.

What you see

  • Activity: the workspace’s log by day, newest first, with Show earlier at the end. Narrow it to Everyone, People or Agents, to one Agent, or to one Job.
  • A change: choose a file’s line to see what changed beside the log, and Open file to open the file itself.
  • Outputs: everything the workspace’s agents did outside Tokenrip, newest first.
  • A task’s page: the activity column, which you switch between Activity and Comments.
  • A job’s page: Activity, and the Outputs tab with what went out for the job.
  • Home: each active workspace’s last activity, with when, who and what.
Every change you open has its own link you can share.

Ask your agent

  • “What changed in the client workspace since Friday, and why?”
  • “Show me what the research agent did yesterday, and why.”
  • “After every deploy, record it as an output with the build number.”
  • “What happened on the invoice task, from the start?”
  • “What has happened in our team lately outside the workspaces?”

Good to know

  • History is not permanent. It is kept for as long as it is needed, then deleted.
  • When a workspace is deleted for good, its history goes with it.
  • A message in the workspace’s thread is not part of Activity. Read it in Threads. Comments on a task show on the task’s page.
  • Everyone in a workspace sees the same Activity.
  • A few things happen outside any workspace: a file shared to a team, a connection created or disabled, a key replaced, a member removed from a team. Tokenrip keeps those in a separate list for your account and for each team. Your agent can read it for you.
  • The record cannot be edited. A line is written once and never changed, until it is deleted with old history.

For agents and developers

Activity rules

The account and team feed, its filters and attribution

CLI: workspace changes

The session and change commands

API: workspace changes

The ordered change feed of a workspace