dsh-quota-pill

dsh-quota-pill

A quota pill for the DeepSeek Harness web GUI: how much is left on every LLM subscription you have, in one collapsible pill next to the composer.

It sits in the composer toolbar, shows the tightest window at a glance, and opens into a per-provider breakdown with reset times. Switch sessions and it follows the model — showing that provider's last reading while it refreshes, then easing into the live number.

What it reads

Provider · Credential DeepSeek (official) · the Platform sign-in DSH already holds — nothing to add Kimi Coding · refs.KIMI_CODING_API_KEY in ~/.dsh/.credentials.yaml OpenCode Go · refs.OPENCODE_GO_API_KEY Command Code · refs.COMMANDCODE_API_KEY or refs.COMMANDCODE_GOAT_API_KEY (DSH names it after its apiKeyEnv) xAI (Grok) · OAuth token from CLIProxyAPI Codex (ChatGPT Plus) · same — and CLIProxyAPI must be running Antigravity · same Open Design (optional) · refs.OPENDESIGN_CONTROL_KEY (vela_ctrl_…) + refs.OPENDESIGN_WORKSPACE_ID; a console session cookie (OPENDESIGN_SESSION_COOKIE) also works as a fallback

DeepSeek works with no setup at all if you are signed in to DeepSeek Platform in DSH — that is the official account, and it is the one most users care about.

The three API-key providers need nothing but the key. The three OAuth providers need CLIProxyAPI installed and logged in, because that is what holds the tokens; this plugin only reads them and never runs an OAuth flow itself.

Every provider is optional, and only the ones you actually have are listed: the pill reads DSH's session/modelCatalog to learn which providers the app can route to, so removing a provider from DSH removes it from the pill too. A provider that is present but has no credential is skipped rather than shown as an empty row.

What "auto-detect" means

• A new key for a supported provider (e.g. adding OPENCODE_GO_API_KEY)

appears on the next All refresh — no DSH restart, no sidecar restart.

• A provider newly added to DSH's catalog appears within the 2-minute catalog

TTL, or immediately when you press Refresh / reopen the All page (a forced read drops the catalog cache). The All page also re-reads when DSH regains focus.

• A brand-new provider type that is not one of the eight adapters needs a

user adapter in ~/.dsh/quota-pill/adapters/ (see below); a catalog entry alone cannot invent its quota API.

Unknown providers and user adapters

If the Current page cannot read the session's provider, it shows a short hint and a 复制提示词 / Copy prompt button. The copied prompt contains the whole adaptation flow; paste it into a new session and an agent can add the provider without touching package files.

User adapters live in:

~/.dsh/quota-pill/adapters/<provider>.mjs

They are outside the npm package, so updating the plugin never overwrites them. The sidecar discovers a new/changed file on the next Refresh (no restart). /setup?fresh=1 reports unadaptedCatalog and userAdapters, and after a plugin update the same check tells the agent which providers still need an adapter. The module contract (exports, read(ctx), standard kinds) is documented in src/user-adapters.js and in AGENTS.md §2b.

Install

DSH Desktop (recommended). Open Plugins → Install, paste a spec, enable the bundle. Pin the version: an unpinned install can resolve 0.2.2, which breaks web boot. 0.2.12 is the first version with every provider, Open Design Control Keys, sidecar auto-start, and user-defined adapters for providers the package does not ship.

@hystudio/dsh-quota-pill@0.2.12

0.2.12 is published on npm. If pnpm's minimumReleaseAge rejects this fresh release, install the branch until it ages:

github:598934321/dsh-quota-pill#master

CLI. If dsh is not on PATH, open DSH Desktop → Settings → Manage dsh Command first. Then:

dsh plugin add @hystudio/dsh-quota-pill@0.2.12

The npm package is @hystudio/... while the repo lives at 598934321/... —

they are the same author under two account names.

The sidecar now starts automatically with the plugin (the host half spawns it when nothing is listening on :8319), so there is no second command for a normal desktop install. If you prefer to manage it yourself, run npx @hystudio/dsh-quota-pill or a service unit — see Running it as a service. The pill says "quota service is not running" only if :8319 is still closed a few seconds after DSH starts.

Then add whichever keys you have:

# ~/.dsh/.credentials.yaml
refs:
  KIMI_CODING_API_KEY: sk-...
  OPENCODE_GO_API_KEY: sk-...
  COMMANDCODE_API_KEY: sk-...

Reload the DSH window. Done.

Let an agent do it

The repo ships AGENTS.md — a step-by-step guide written for an AI agent. Ask your agent to set the plugin up and it will detect what you already have, ask only for what it cannot know (your keys), write them safely, and verify each provider end to end.

Read AGENTS.md in this repo and set up dsh-quota-pill for me.

Running it as a service

The host plugin already starts the sidecar; this section is only for users who want it independent of DSH (server, headless, multiple app profiles). Set QUOTA_PILL_NO_AUTOSTART=1 so the plugin does not race the service.

npx in a terminal dies with the terminal. Pick one:

macOS (launchd) — ~/Library/LaunchAgents/dev.quota-pill.plist

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN"
  "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0"><dict>
  <key>Label</key><string>dev.quota-pill</string>
  <key>ProgramArguments</key>
  <array>
    <string>/usr/bin/env</string><string>node</string>
    <string>/absolute/path/to/bin/quota-pill-server.js</string>
  </array>
  <key>RunAtLoad</key><true/>
  <key>KeepAlive</key><true/>
</dict></plist>

launchctl load ~/Library/LaunchAgents/dev.quota-pill.plist

Linux (systemd) — ~/.config/systemd/user/quota-pill.service

[Unit]
Description=dsh-quota-pill sidecar

[Service]
ExecStart=/usr/bin/env node /absolute/path/to/bin/quota-pill-server.js
Restart=always

[Install]
WantedBy=default.target

systemctl --user enable --now quota-pill

How it works

DSH web GUI
   └─ client.js          renders the pill; never sees a credential
        │  HTTP (localhost)
        ▼
      sidecar           reads ~/.dsh/.credentials.yaml and ~/.cli-proxy-api/*.json,
                        calls each provider's quota endpoint, caches 60s

Two processes because credentials must never reach the browser, and because upstream calls need real timeouts and a TLS stack Node's fetch does not have for some hosts.

Endpoints

Path · Purpose GET /health · liveness + provider list GET /quota?model= · one provider's reading, selected by model prefix GET /quota/all · every provider, concurrently GET /setup · what each provider still needs

?fresh=1 bypasses the 60-second cache.

Response

{
  "kind": "subscription-quota",   // subscription-quota | payg | none | probe-failed
  "provider": "opencode",
  "displayName": "OpenCode Go",
  "remainingPct": 17,             // remaining, never used
  "window": "monthly",            // the binding window
  "resetAt": "2026-11-12T00:00:00.000Z",
  "products": { "rolling": 9, "weekly": 67, "monthly": 83 },  // per-window used%
  "windows": [                    // one row each in the UI
    { "key": "monthly", "remainPct": 17, "resetAt": "..." }
  ],
  "cached": false,
  "at": 1792500000000
}

kind is meant to be exhaustive so the UI never has to guess:

• subscription-quota — a real percentage exists
• payg — metered billing, no ceiling (shows spend, not a percentage)
• none — not configured; error names the missing credential
• probe-failed — configured but the upstream call failed

Notes on specific providers

Codex is read through CLIProxyAPI's management API, which reports the X-Codex-* headers CLIProxyAPI observed while proxying real traffic. The direct chatgpt.com/backend-api/codex/usage endpoint sits behind Cloudflare and rejects non-browser TLS fingerprints, so it only works as a fallback. CLIProxyAPI must be *running*, not merely installed.

xAI has three account generations: grandfathered plans report a percentage directly, current plans only reveal allowance in x-ratelimit-* response headers of a one-token completion, and unified-billing accounts have no ceiling at all (reported as payg).

Antigravity shares one quota pool across all Gemini models, so it reports a single aggregate row — per-model rows would all be the same number and the label would flicker between models you do not even have configured.

Command Code does not expose a monthly ceiling. The adapter derives it from monthlyCap = weekly.cap × 2, an identity verified against two different plans (31.6 + 38.4 = 70 with weekly.cap 35; 0.225 + 9.775 = 10 with weekly.cap 6).

Open Design keeps its allowance in the cloud control plane, not in the sk-amr-… inference key. The preferred credential is a Control Key (vela_ctrl_…), which is accepted by the Bearer-only /billing/preflight and /wallet/balance endpoints alongside x-vela-workspace-id. Get one with the bundled device sign-in (the same flow as the official vela login):

npx -p @hystudio/dsh-quota-pill dsh-quota-pill-opendesign-login --write

It prints an activation URL, waits for you to approve it in a signed-in browser, then writes OPENDESIGN_CONTROL_KEY and OPENDESIGN_WORKSPACE_ID into ~/.dsh/.credentials.yaml. The workspace id is optional when using a Control Key — it is otherwise resolved from /workspaces/current.

The older console-session route (__Secure-better-auth.session_token cookie, browser-session-only /billing/coding-plan-usage) is still supported as a fallback; that cookie is HttpOnly, so copy it from devtools → Application → Cookies. Both routes go through system curl because Cloudflare rejects Node's TLS fingerprint on amr-api with error code: 1010.

DeepSeek is a prepaid CNY wallet, so there is no percentage to show — the pill shows the balance and, separately, any bonus credit. Two traps worth recording: the stored credential is a *Platform login grant*, not an api.deepseek.com API key (sending it there returns 401), and calling platform.deepseek.com directly returns 429 from Cloudflare. The balance is therefore read through DSH's own account/getBalance RPC, which already holds the grant and projects the response down to wallet currency and balance. We never handle the Platform token ourselves.

Privacy

• Credentials are read from disk, used for the upstream call, and never logged.
• The sidecar binds 127.0.0.1 only.
• Nothing is sent anywhere except the provider's own quota endpoint.
• The client shows no credential, and the sidecar's JSON carries no token.

Development

git clone https://github.com/598934321/dsh-quota-pill
cd dsh-quota-pill
npm test              # 70 tests, offline
npm start             # sidecar on :8319

Tests are offline by design: they point the credential layer at a nonexistent directory so they never touch a live account. Fixtures in test/fixtures.js are synthetic — no real emails, account ids, or balances.

License

MIT