pi-statusline

A rich custom status bar for pi that replaces the default footer. Provider-aware: shows live account balance for DeepSeek and rolling quota for GLM/ZAI, plus session tokens, cost, context window, and timing.

Features

• Provider-aware usage — special support for DeepSeek and GLM/ZAI (see Provider Support)
• DeepSeek peak/off-peak pricing — session cost follows DeepSeek's 2026-08-17 peak/off-peak billing (peak 9:00-12:00 / 14:00-18:00 Beijing time on weekdays; weekends are off-peak all day since 2026-08-23), priced per-message by timestamp, with a live peak/off-peak indicator in the footer (see DeepSeek)
• Token usage: input, output, cache read/write, total per session, with cache hit rate
• Cost: cumulative session cost with currency auto-detection (¥ for DeepSeek/CNY, $ otherwise)
• Context window: usage percentage and size
• Timing: elapsed time + tokens/sec for last response
• Git branch: current branch shown in cwd display

Provider Support

This extension uses a provider-aware strategy: when the active model belongs to a supported provider, the status bar shows live account-level usage data fetched directly from that provider's API. Other providers fall back to session-scoped cost only.

DeepSeek

Shows live account balance, weekly token usage, and peak/off-peak aware session cost. Since DeepSeek introduced peak/off-peak pricing on 2026-08-17, pi-statusline tracks it automatically.

Peak / Off-peak pricing (since 2026-08-17)

DeepSeek now bills in two periods — peak hours are 9:00-12:00 and 14:00-18:00 Beijing time (UTC+8) on weekdays, all other hours are off-peak, and off-peak is half the peak price. Since 2026-08-23 00:00 Beijing time, weekends (Saturday/Sunday) no longer distinguish periods and bill at the off-peak price all day. Official CNY prices per million tokens:

Model	Period	Cache-hit input	Cache-miss input	Output
deepseek-flash	off-peak / peak	¥0.02 / ¥0.04	¥1 / ¥2	¥4 / ¥8
deepseek-v4-pro	off-peak / peak	¥0.15 / ¥0.30	¥4.5 / ¥9.0	¥13.5 / ¥27.0

• Legacy model ids — deepseek-v4-flash and deepseek-v4-flash-vision-exp are retired but still callable; requests are served by DeepSeek-V4.1-Flash and billed at the Flash rates, so they share the deepseek-flash prices.

• Per-message pricing by timestamp — each message is priced at the rate of its own time (AssistantMessage.timestamp), so a long session spanning a period boundary (9:00/12:00/14:00/18:00) accounts every message correctly instead of using a single fixed rate.

• Peak/off-peak indicator — the footer appends the current period so you always know which rate is in effect: peak during weekday 9:00-12:00 / 14:00-18:00 Beijing time, off-peak otherwise (including all day on weekends since 2026-08-23):

  ¥0.12/50.00 · 7d:1.2M · off-peak
  ¥0.12/50.00 · 7d:1.2M · peak
  

• Registry pricing patch — for CNY-billed accounts the model registry cost is aligned to the current period at runtime (and re-aligned automatically when the period flips), so pi's recorded usage.cost tracks the period too.

• USD-billed accounts keep pi's built-in USD prices (no CNY override, no peak/off-peak indicator).

Balance & weekly tokens

Segment	Example	Source
Balance	¥0.12/50.00	GET /user/balance (cached 5 min)
Weekly tokens	7d:1.2M	Local session file scan (rolling 7 days)

• Currency auto-detected as ¥ (CNY) from the balance API response
• Balance is fetched on startup and refreshed in the background; cached for 5 minutes
• Triggered when model.provider === "deepseek"

GLM / ZAI

Shows 5-hour rolling quota and weekly quota (or natural-week usage as fallback).

Segment	Example	Source
5h quota	Usage 42%(1h23m)	GET /api/monitor/usage/quota/limit (unit:3)
Weekly quota	W:35%(1.2M,3d4h)	quota/limit (unit:6) + model-usage API
Natural week	W:1.2M	model-usage API (Mon 00:00 local → now)

• 5-hour rolling window: percentage used + countdown to reset
• Weekly quota: if the plan exposes a unit:6 weekly limit, shows percentage, tokens, and reset countdown; otherwise falls back to real usage for the current natural week (host local timezone)
• Account level: also fetched (visible in /status-debug)
• Supports both zai (api.z.ai) and zai-coding-cn (open.bigmodel.cn) endpoints

Other providers

Session-scoped cost only ($0.12 or ¥0.12), with no live account data.

Install

Requires the pi CLI.

From npm (recommended)

# Global (user) install — available in every project
pi install @fyeeme/pi-statusline

# Project-local — written to .pi/settings.json, shareable with your team
pi install -l @fyeeme/pi-statusline

# Pinned version — skipped by `pi update`
pi install @fyeeme/pi-statusline@1.0.2

# Try it once without saving (current run only)
pi -e @fyeeme/pi-statusline

From GitHub

Source: fyeeme/pi-packages.

# HTTPS shorthand
pi install git:github.com/fyeeme/pi-packages
# Pin to a tag or commit (skipped by `pi update`)
pi install git:github.com/fyeeme/pi-packages@v1.0.2
# Raw URL form
pi install https://github.com/fyeeme/pi-packages

See the Pi Packages guide on pi.dev for the full list of source types, scopes, and pi update behavior.

Commands

Command	Description
/currency [auto|¥|$]	Toggle cost currency display (auto follows DeepSeek → ¥)
/status-debug	Dump session stats + provider usage to /tmp/pi-status-debug.log

Status Bar Layout

~/projects/my-repo (main)                    deepseek-v4-pro · xhigh
tokens 65k(in 12k, out 8k, cache 45k) · ¥0.12/50.00 · 7d:1.2M · off-peak · 45.2%/64k · 2m30s 38.2tok/s

Line 1: cwd + git branch (left) | model + thinking level (right)

Line 2: token stats · provider usage (incl. DeepSeek peak/off-peak indicator) · context window · timing

GLM/ZAI example

~/projects/my-repo (main)                       glm-4.6 · high
tokens 65k(in 12k, out 8k, cache 45k) · Usage 42%(1h23m) · W:35%(1.2M,3d4h) · 45.2%/64k · 2m30s 38.2tok/s

Environment Variables

None required. API keys are resolved automatically by pi's model registry for the active provider.

How It Works

The extension registers a UsageProvider per supported provider and refreshes usage data in the background (on startup, on model switch, and after each agent run). Results are cached with a 5-minute TTL (DeepSeek) or per-refresh (ZAI) to avoid hammering the provider APIs.

All network calls are best-effort with a 5-second timeout — failures stay silent and never block the status bar.

Platform API note: the custom footer relies on ctx.ui.setFooter's third parameter — footerData: ReadonlyFooterDataProvider (for getGitBranch/getExtensionStatuses/onBranchChange) — and on returning dispose from the footer factory. These are part of the 0.84.x ExtensionAPI surface (see dist/core/extensions/types.d.ts), but pi's docs/extensions.md currently documents only the (tui, theme) form. If a future pi release changes the setFooter signature, check this extension's footer wiring first.

Changelog

See CHANGELOG.md.

License

MIT