Displayed data
The status line provides 26 configurable fields. Use the interactive preview to choose fields, order, and colors. The v0.1.2 marker at the right is the plugin version; codex_version is the Codex CLI version.
All 26 fields
| Field | Meaning |
|---|---|
label | Fixed label, default CODEX |
model | Current model |
reasoning | Reasoning effort such as high |
activity | Review, Plan, or Goal state |
provider | Model provider |
personality | Session personality setting |
service_tier | Service tier |
cwd | Working directory |
project | Git root name, omitted outside Git |
git | Current Git branch |
permissions | Sandbox permission summary |
approval | Approval policy |
context | Remaining context percentage or meter |
context_used | Used context percentage |
context_window | Context capacity, such as 272K window |
used_tokens | Cumulative input and output, such as ↑10M ↓204K |
cache_rate | Cached input divided by total input |
cost | Estimated cost, such as Cost ~$22.48 |
input_tokens | Cumulative input including cached input |
cached_tokens | Cumulative cached input |
output_tokens | Cumulative output including reasoning output |
reasoning_tokens | Cumulative reasoning output |
thread_id | First eight characters of the session ID |
task_progress | Preview-only task progress |
codex_version | Codex CLI version |
icon | Project icon, fixed before the plugin version |
Quota, PR, and Enterprise credit data are outside the extension's scope.
Special states
activity displays REVIEW, PLAN, or a Goal state. The priority is REVIEW, then PLAN, then Goal. Goal states include active, paused, blocked, usage limited, budget limited, and complete. Labels can be changed with activity.labels.
Width and rows
| Columns | Behavior |
|---|---|
| Under 60 | Keep label, model, activity, context, and the right marker |
| 60–89 | Hide advanced metadata and token details |
| 90–119 | Use an eight-cell context meter |
| 120 or more | Use a ten-cell meter and show more fields |
When space is still short, lower-priority fields are hidden. Two-row layouts put context, tokens, cache rate, and cost on the second row.
Tokens and cost
used_tokens shows cumulative input and output. Input includes cached input, and output includes reasoning output. cache_rate is cached input divided by total input. A missing input shows Cache —; a nonzero input with no cache hit shows Cache 0%.
Cost is estimated per request and then summed:
request cost = ((input - cached input) × input price
+ cached input × cached price
+ output × output price) / 1,000,000
session cost = sum of request costs
Prices are USD per million tokens. Switching models does not change the price assigned to earlier requests. Incomplete history or missing model pricing yields Cost —. The estimate excludes long-context, acceleration, cache-write, and tool charges.
Remaining context
The context fields use the latest request, while token totals cover the whole session. The remaining percentage uses Codex's 12,000-token baseline, is clamped to 0–100%, and falls back to the current context token count when capacity is unknown.
Waiting states
| Display | Meaning |
|---|---|
waiting | No session record has been associated with the current pane |
tokens: waiting | A session is associated, but no usage data has arrived |
A new session normally needs one message first. See waiting-state troubleshooting.
Refresh rate
The plugin checks process and session changes every second and Git information every five seconds. WezTerm's status_update_interval also affects the visible refresh rate.