一个 MCP 服务器（stdio），将你的 OpenCode 订阅数据以工具的形式暴露出来——你的 Go 配额和 Zen 预付余额——供任何支持 MCP 的客户端（pi、Claude Desktop、opencode、Cursor 等）使用。
query_go_usage — OpenCode Go 配额，涵盖滚动（5 小时）、每周和每月窗口。获取 /workspace/{id}/go 页面并解析内嵌的 RSC 负载。
query_zen_balance — OpenCode Zen 预付余额。获取工作区仪表板页面并解析内嵌的 SolidJS SSR 标量。
请求在瞬时故障（5xx、429——遵循 Retry-After）时以退避方式重试，并在 15 秒后超时。
零依赖。Node.js ≥ 18。
npx -y opencode-balance-mcp \
--workspace-id wrk_xxx \
--auth-cookie "Fe26.2**..."Related MCP server: kitepass-mcp
服务器需要你的 OpenCode 工作区 ID 和 认证 cookie。优先级顺序：
CLI 参数：--workspace-id / --auth-cookie
环境变量：OPENCODE_GO_WORKSPACE_ID / OPENCODE_GO_AUTH_COOKIE
⚠️
authCookie是你的 OpenCode 会话 cookie（以Fe26.2**开头）。它会过期——通过登录 https://opencode.ai → DevTools → Application → Cookies →auth来刷新它。
query_go_usage无参数。返回所有三个窗口的 Go 订阅配额：
{
"timestamp": "...",
"workspaceId": "wrk_...",
"rolling": { "status": "ok", "usagePercent": 1, "limitUsd": 12, "estSpentUsdLow": 0.12, "estSpentUsdHigh": 0.24, "resetsInSeconds": 8660, "resetsIn": "2h 24m" },
"weekly": { "status": "ok", "usagePercent": 1, "limitUsd": 30, "...": "..." },
"monthly": { "status": "ok", "usagePercent": 0, "limitUsd": 60, "...": "..." }
}usagePercent 是一个整数——0% 仅表示使用了窗口限额的不到 1%（5 小时 $12 / 每周 $30 / 每月 $60）。参见 opencode.ai/docs/go。
estSpentUsdLow/High 给出了当前使用百分比下估算花费的范围。
resetsIn 是从 resetsInSeconds 派生的人类可读倒计时。
query_zen_balance无参数。读取工作区仪表板页面并解析 Zen 余额（SolidJS SSR 水合数据）：
{
"timestamp": "...",
"workspaceId": "wrk_...",
"plan": "pay-as-you-go",
"balanceRaw": 0,
"balanceUsd": 0,
"balanceFormatted": "$0.00",
"autoReload": { "triggerUsd": 5, "triggerMinUsd": 5, "reloadUsd": 20, "reloadMinUsd": 10 },
"reloadAmount": 20,
"reloadTrigger": 5
}balanceRaw 是一个以 1e-8 美元单位 表示的整数（与仪表板客户端使用的单位相同：formatBalance = amount / 1e8）。负值 = 预付信用，正值 = 欠款金额。
balanceUsd / balanceFormatted 由它派生；符号在标签中显示为 (credit) / (owed)。
autoReload 反映 Zen 的自动充值：当余额低于 triggerUsd 时，添加 reloadUsd。
没有官方的 Zen 余额 API，因此服务器以与探测 Go 配额相同的方式抓取仪表板。解析器位于
src/parse.mjs中，如果页面结构发生变化，可能需要更新。
~/.config/opencode/opencode.json）{
"mcp": {
"opencode-balance": {
"type": "local",
"command": ["npx", "-y", "opencode-balance-mcp"],
"environment": {
"OPENCODE_GO_WORKSPACE_ID": "wrk_xxx",
"OPENCODE_GO_AUTH_COOKIE": "Fe26.2**..."
}
}
}
}{
"mcpServers": {
"opencode-balance": {
"command": "npx",
"args": ["-y", "opencode-balance-mcp"],
"env": {
"OPENCODE_GO_WORKSPACE_ID": "wrk_xxx",
"OPENCODE_GO_AUTH_COOKIE": "Fe26.2**..."
}
}
}
}.mcp.json）{
"mcpServers": {
"opencode-balance": {
"command": "npx",
"args": ["-y", "opencode-balance-mcp"],
"env": {
"OPENCODE_GO_WORKSPACE_ID": "wrk_xxx",
"OPENCODE_GO_AUTH_COOKIE": "Fe26.2**..."
}
}
}
}或通过 CLI：claude mcp add opencode-balance -e OPENCODE_GO_WORKSPACE_ID=wrk_xxx -e OPENCODE_GO_AUTH_COOKIE="Fe26.2**..." -- npx -y opencode-balance-mcp
~/.codex/config.toml）[mcp_servers.opencode-balance]
command = "npx"
args = ["-y", "opencode-balance-mcp"]
enabled = true
[mcp_servers.opencode-balance.env]
OPENCODE_GO_WORKSPACE_ID = "wrk_xxx"
OPENCODE_GO_AUTH_COOKIE = "Fe26.2**..."~/.pi/agent/mcp.json）{
"mcpServers": {
"opencode-balance": {
"command": "npx",
"args": ["-y", "opencode-balance-mcp"],
"env": {
"OPENCODE_GO_WORKSPACE_ID": "wrk_xxx",
"OPENCODE_GO_AUTH_COOKIE": "Fe26.2**..."
}
}
}
}npm start # run the server directly
npm test # unit tests (parser, zen formatting, HTTP retry policy) + a protocol smoke test协议冒烟测试（test/test.mjs）通过 stdio 启动服务器，并验证 initialize、tools/list、tools/call 以及 tools/call 的错误处理。在没有凭据的情况下，它会按预期以 isError 结果失败两次工具调用。
src/index.mjs — MCP stdio 协议、凭据（CLI/环境变量）、页面获取粘合代码
src/http.mjs — 网络策略：超时、带退避的重试、429/Retry-After 处理
src/parse.mjs — 纯解析/格式化辅助函数（无 I/O，直接进行单元测试）
npm publishMIT
Query OpenCode Go subscription quota: rolling (5h) / weekly / monthly usage percent, estimated USD spent and reset countdown. No arguments needed; credentials are provided at startup via CLI args (--workspace-id / --auth-cookie) or env vars.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It adds value by disclosing that credentials are pre-configured via CLI args or env vars, which prevents unnecessary auth handling. However, it does not mention potential error conditions, whether the action is read-only (though implied by 'query'), or any rate limits. It neither contradicts nor fully discloses behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences with no wasted words. The first sentence states purpose and outputs; the second clarifies argument and credential requirements. Information is front-loaded and every word serves a purpose.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a zero-parameter query tool with no output schema, the description covers the essential details: what it returns and that no arguments are needed. It does not specify the exact format of the reset countdown (e.g., seconds, minutes), which could cause minor ambiguity, but overall the tool is simple enough that the description is nearly complete.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters, so baseline is 4. The description reinforces this by explicitly stating 'No arguments needed,' which adds clarity beyond the empty schema. Since there are no parameters to document, this is fully adequate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool queries OpenCode Go subscription quota and enumerates specific outputs (usage percent, estimated USD spent, reset countdown). It uses a specific verb 'query' and resource 'Go subscription quota,' making it distinguishable from the sibling tool query_zen_balance.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It explicitly states 'No arguments needed' and explains how credentials are provided, which tells the agent it doesn't need to pass authentication. It does not explicitly name the sibling or contrast usage, but the purpose is clear enough that the agent can infer when to use it. Lacks an explicit 'when not to use' but provides sufficient context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Query OpenCode Zen prepaid balance (pay-as-you-go): balance in 1e-8 USD units (negative = credit), formatted USD, and auto-reload settings. No arguments needed; credentials are provided at startup via CLI args (--workspace-id / --auth-cookie) or env vars.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden and discloses useful behavioral details: balance units in 1e-8 USD, negative values meaning credit, formatted USD output, and auto-reload settings. It also clarifies credential sourcing, which is important operational behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two efficiently written sentences convey the purpose, output, required arguments, and credential mechanism with no filler. Every clause earns its place and the core subject is front-loaded.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a zero-parameter read-only query tool with no output schema, the description is complete: it names the resource, details the returned data, explains the unit semantics, states no arguments are needed, and tells where credentials come from. Nothing necessary for correct invocation is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are zero parameters and schema coverage is 100%, so the baseline is 4. The description reinforces that no arguments are needed and goes further by explaining how authentication is already handled, which adds meaning beyond the empty schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description names a specific verb (Query), a specific resource (OpenCode Zen prepaid balance), and the scope (pay-as-you-go balance). It lists the exact data returned, which distinguishes it from the sibling query_go_usage without needing to compare schemas.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly states that no arguments are needed and explains how credentials are supplied, which is essential contextual guidance for invoking the tool. It does not explicitly name the sibling as an alternative, but the resource and data scope make the intended use clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
The two tools query distinct resources: one for Go subscription quota and one for Zen prepaid balance. Their descriptions clearly differentiate the data returned (usage percentages vs. balance in USD units), making confusion unlikely.
Both tool names follow the exact same pattern: 'query_' prefix followed by the resource type ('go_usage' and 'zen_balance'). This is fully consistent and predictable.
With only two tools, the surface is minimal, but the server's scope is narrowly defined as querying two distinct balance types. The count feels appropriate given the focused purpose, though slightly thin if broadenability was expected.
The server covers the two core balance query operations for its domain. Missing a combined query or historical data are minor gaps that agents can work around, but the essential read operations are present.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
A paid remote MCP for AI SDK data query MCP, built to return verdicts, receipts, usage logs, and aud
A paid remote MCP for OpenAI Codex agent coordination MCP, built to return verdicts, receipts, usage
Query OneLens cloud-cost data in natural language: breakdowns, trends, cost centers. Read-only.
Governed MCP gateway: one endpoint for your tools, with credential custody and audit log.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/gralliry/opencode-balance-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server