Context Breakdown

The Context Breakdown page shows you exactly how your agent’s context window is being consumed — step by step, broken down by component. Use it to understand why a session is running out of context, which component is growing the fastest, and where you might be able to reduce token usage.

This page requires cache trace data to be enabled. If cache trace is not active, the page displays an “Unavailable” banner with instructions on how to enable it.

Enabling cache trace

The cache trace status is shown as a colored dot at the top of the page: green when enabled, gray when not. Two ways to enable it: Option 1 — add to ~/.openclaw/openclaw.json:

{
  "diagnostics": {
    "cacheTrace": {
      "enabled": true,
      "includeMessages": true,
      "includePrompt": true,
      "includeSystem": true
    }
  }
}

Option 2 — send the keyword OPENCLAW_CACHE_TRACE to OpenClaw and let it enable the setting for you.

The cache trace file grows quickly. Set up a cron job to clean it periodically, or it will consume significant disk space over time.

Selecting a session

Use the session picker to choose which session to inspect. The dropdown shows the session ID, agent name, model, cost, date, and number of steps for each session. Type to search by any of these fields. Click × to clear the current selection and return to the empty state.

Per-step cost chart

A bar chart showing the token cost at each step of the session. Bars are color-coded by relative cost:

Color	Meaning
Red	≥80% of the maximum step cost — the most expensive steps
Amber	≥50% of the maximum step cost
Blue	Below 50% — normal cost

Hover over any bar to see the token breakdown: input, output, cache write, and cache read.

Context composition chart

A stacked bar chart showing how the context window is composed at each step. Three components are tracked:

Color	Component	Description
Blue	System	The system prompt and instructions
Cyan	History	Conversation history — previous messages between the user and agent
Amber	Tools	Tool call results and tool-related content

This chart tells you at a glance which component is dominating the context window. If the amber (Tools) section grows rapidly, your agent is accumulating large tool outputs. If cyan (History) grows steadily, the conversation itself is consuming the space.

Per-turn table

A detailed table breaking down the context window at each step:

Column	Description
Step	The step number in the session
System	Tokens consumed by the system prompt
History	Tokens consumed by conversation history
Tools	Tokens consumed by tool results
Total	Total tokens used at this step
Pressure	Context utilization as a percentage of the model’s context limit
Cost	Token cost for this step (shown only if cost data is available)

Column headers have tooltips explaining what each value means. The pressure column tooltip shows the model’s context capacity (e.g. “200K”).

Get Started

Monitoring

Debugging

Security Audit

Engineering Note

Reference

Enabling cache trace

Selecting a session

Per-step cost chart

Context composition chart

Per-turn table

​Enabling cache trace

​Selecting a session

​Per-step cost chart

​Context composition chart

​Per-turn table

Enabling cache trace

Selecting a session

Per-step cost chart

Context composition chart

Per-turn table