Data Models: Auth File Structures & API Response Formats

💡 This appendix is for developers: If you want to understand how the plugin reads and parses authentication files, or if you want to extend support for more platforms yourself, this provides a complete data model reference.

What You'll Learn

Understand which authentication files the plugin reads
Understand the API response formats for each platform
Know how to extend the plugin to support new platforms

Appendix Contents

Auth file structures (3 config files)
API response formats (5 platforms)
Internal data types

Auth File Structures

Main Auth File: `~/.local/share/opencode/auth.json`

OpenCode official authentication storage. The plugin reads authentication information for OpenAI, Zhipu AI, Z.ai, and GitHub Copilot from here.

typescript

interface AuthData {
  /** OpenAI OAuth authentication */
  openai?: OpenAIAuthData;

  /** Zhipu AI API authentication */
  "zhipuai-coding-plan"?: ZhipuAuthData;

  /** Z.ai API authentication */
  "zai-coding-plan"?: ZhipuAuthData;

  /** GitHub Copilot OAuth authentication */
  "github-copilot"?: CopilotAuthData;
}

OpenAI Auth Data

typescript

interface OpenAIAuthData {
  type: string;        // Fixed value "oauth"
  access?: string;     // OAuth Access Token
  refresh?: string;    // OAuth Refresh Token
  expires?: number;    // Expiration timestamp (milliseconds)
}

Data Source: OpenCode official OAuth authentication flow

Zhipu AI / Z.ai Auth Data

typescript

interface ZhipuAuthData {
  type: string;   // Fixed value "api"
  key?: string;    // API Key
}

Data Source: API Key configured by user in OpenCode

GitHub Copilot Auth Data

typescript

interface CopilotAuthData {
  type: string;        // Fixed value "oauth"
  refresh?: string;     // OAuth Token
  access?: string;      // Copilot Session Token (optional)
  expires?: number;    // Expiration timestamp (milliseconds)
}

Data Source: OpenCode official OAuth authentication flow

Copilot PAT Configuration: `~/.config/opencode/copilot-quota-token.json`

Optional Fine-grained PAT (Personal Access Token) configuration for querying quotas through GitHub public API (no Copilot permissions required).

typescript

interface CopilotQuotaConfig {
  /** Fine-grained PAT (requires "Plan" read permission) */
  token: string;

  /** GitHub username (required for API calls) */
  username: string;

  /** Copilot subscription type (determines monthly quota limit) */
  tier: CopilotTier;
}

/** Copilot subscription type enum */
type CopilotTier = "free" | "pro" | "pro+" | "business" | "enterprise";

Quota limits for each subscription type:

tier	Monthly Limit (Premium Requests)
`free`	50
`pro`	300
`pro+`	1,500
`business`	300
`enterprise`	1,000

Google Cloud Accounts: `~/.config/opencode/antigravity-accounts.json`

Account file created by opencode-antigravity-auth plugin, supporting multiple accounts.

typescript

interface AntigravityAccountsFile {
  version: number;               // File format version
  accounts: AntigravityAccount[];
}

interface AntigravityAccount {
  /** Google email (for display) */
  email?: string;

  /** OAuth Refresh Token (required) */
  refreshToken: string;

  /** Google project ID (one of two) */
  projectId?: string;

  /** Managed project ID (one of two) */
  managedProjectId?: string;

  /** Account added timestamp (milliseconds) */
  addedAt: number;

  /** Last used timestamp (milliseconds) */
  lastUsed: number;

  /** Rate limit reset times for each model (model key → timestamp) */
  rateLimitResetTimes?: Record<string, number>;
}

Data Source: OAuth authentication flow from opencode-antigravity-auth plugin

API Response Formats

OpenAI Response Format

API Endpoint: GET https://chatgpt.com/backend-api/wham/usage

Authentication Method: Bearer Token (OAuth Access Token)

typescript

interface OpenAIUsageResponse {
  /** Plan type: plus, team, pro, etc. */
  plan_type: string;

  /** Quota limit information */
  rate_limit: {
    /** Whether limit is reached */
    limit_reached: boolean;

    /** Primary window (3 hours) */
    primary_window: RateLimitWindow;

    /** Secondary window (24 hours, optional) */
    secondary_window: RateLimitWindow | null;
  } | null;
}

/** Limit window information */
interface RateLimitWindow {
  /** Used percentage */
  used_percent: number;

  /** Limit window duration (seconds) */
  limit_window_seconds: number;

  /** Seconds until reset */
  reset_after_seconds: number;
}

Example Response:

json

{
  "plan_type": "team",
  "rate_limit": {
    "limit_reached": false,
    "primary_window": {
      "used_percent": 15,
      "limit_window_seconds": 10800,
      "reset_after_seconds": 9000
    },
    "secondary_window": {
      "used_percent": 23,
      "limit_window_seconds": 86400,
      "reset_after_seconds": 43200
    }
  }
}

Zhipu AI / Z.ai Response Format

API Endpoints:

Zhipu AI: GET https://bigmodel.cn/api/monitor/usage/quota/limit
Z.ai: GET https://api.z.ai/api/monitor/usage/quota/limit

Authentication Method: Authorization Header (API Key)

typescript

interface QuotaLimitResponse {
  code: number;   // 200 on success
  msg: string;    // Error message ("success" on success)
  data: {
    limits: UsageLimitItem[];
  };
  success: boolean;
}

/** Single limit item */
interface UsageLimitItem {
  /** Limit type */
  type: "TOKENS_LIMIT" | "TIME_LIMIT";

  /** Current value */
  currentValue: number;

  /** Total value */
  usage: number;

  /** Usage percentage */
  percentage: number;

  /** Next reset timestamp (milliseconds, only for TOKENS_LIMIT) */
  nextResetTime?: number;
}

Limit type explanations:

type	Description	Reset Cycle
`TOKENS_LIMIT`	5-hour Token limit	5 hours
`TIME_LIMIT`	MCP monthly quota	1 month

Example Response:

json

{
  "code": 200,
  "msg": "success",
  "success": true,
  "data": {
    "limits": [
      {
        "type": "TOKENS_LIMIT",
        "currentValue": 500000,
        "usage": 10000000,
        "percentage": 5,
        "nextResetTime": 1737926400000
      },
      {
        "type": "TIME_LIMIT",
        "currentValue": 120,
        "usage": 2000,
        "percentage": 6
      }
    ]
  }
}

GitHub Copilot Response Format

Copilot supports two API query methods with different response formats.

Method 1: Internal API (requires Copilot permission)

API Endpoint: GET https://api.github.com/copilot_internal/user

Authentication Method: Bearer Token (Token after OAuth or Token Exchange)

typescript

interface CopilotUsageResponse {
  /** SKU type (for distinguishing subscriptions) */
  access_type_sku: string;

  /** Analytics tracking ID */
  analytics_tracking_id: string;

  /** Assignment date */
  assigned_date: string;

  /** Whether can sign up for limited plan */
  can_signup_for_limited: boolean;

  /** Whether chat is enabled */
  chat_enabled: boolean;

  /** Copilot plan type */
  copilot_plan: string;

  /** Quota reset date (format: YYYY-MM) */
  quota_reset_date: string;

  /** Quota snapshots */
  quota_snapshots: QuotaSnapshots;
}

/** Quota snapshots */
interface QuotaSnapshots {
  /** Chat quota (optional) */
  chat?: QuotaDetail;

  /** Completions quota (optional) */
  completions?: QuotaDetail;

  /** Premium Interactions (required) */
  premium_interactions: QuotaDetail;
}

/** Quota detail */
interface QuotaDetail {
  /** Quota entitlement */
  entitlement: number;

  /** Overage count */
  overage_count: number;

  /** Whether overage is permitted */
  overage_permitted: boolean;

  /** Percentage remaining */
  percent_remaining: number;

  /** Quota ID */
  quota_id: string;

  /** Quota remaining */
  quota_remaining: number;

  /** Remaining count (same as quota_remaining) */
  remaining: number;

  /** Whether unlimited */
  unlimited: boolean;
}

Method 2: Public Billing API (requires Fine-grained PAT)

API Endpoint: GET https://api.github.com/users/{username}/settings/billing/premium_request/usage

Authentication Method: Bearer Token (Fine-grained PAT, requires "Plan" read permission)

typescript

interface BillingUsageResponse {
  /** Time period */
  timePeriod: {
    year: number;
    month?: number;
  };

  /** Username */
  user: string;

  /** Usage items list */
  usageItems: BillingUsageItem[];
}

/** Usage item */
interface BillingUsageItem {
  /** Product name */
  product: string;

  /** SKU identifier */
  sku: string;

  /** Model name (optional) */
  model?: string;

  /** Unit type (e.g., "requests") */
  unitType: string;

  /** Total requests (before discount) */
  grossQuantity: number;

  /** Net requests (after discount) */
  netQuantity: number;

  /** Quota limit (optional) */
  limit?: number;
}

Example Response:

json

{
  "timePeriod": {
    "year": 2026,
    "month": 1
  },
  "user": "octocat",
  "usageItems": [
    {
      "product": "GitHub Copilot",
      "sku": "Copilot Premium Request",
      "model": "gpt-4o",
      "unitType": "requests",
      "grossQuantity": 229,
      "netQuantity": 229,
      "limit": 300
    },
    {
      "product": "GitHub Copilot",
      "sku": "Copilot Premium Request",
      "model": "claude-3-5-sonnet",
      "unitType": "requests",
      "grossQuantity": 71,
      "netQuantity": 71,
      "limit": 300
    }
  ]
}

Google Cloud Response Format

API Endpoint: POST https://cloudcode-pa.googleapis.com/v1internal:fetchAvailableModels

Authentication Method: Bearer Token (OAuth Access Token)

Request Body:

json

{
  "project": "your-project-id"
}

typescript

interface GoogleQuotaResponse {
  /** Model list (key is model ID) */
  models: Record<
    string,
    {
      /** Quota information (optional) */
      quotaInfo?: {
        /** Remaining fraction (0-1) */
        remainingFraction?: number;

        /** Reset time (ISO 8601 format) */
        resetTime?: string;
      };
    }
  >;
}

Example Response:

json

{
  "models": {
    "gemini-3-pro-high": {
      "quotaInfo": {
        "remainingFraction": 0.83,
        "resetTime": "2026-01-23T20:00:00Z"
      }
    },
    "gemini-3-pro-image": {
      "quotaInfo": {
        "remainingFraction": 0.91,
        "resetTime": "2026-01-23T20:00:00Z"
      }
    },
    "gemini-3-flash": {
      "quotaInfo": {
        "remainingFraction": 1.0,
        "resetTime": "2026-01-23T20:00:00Z"
      }
    },
    "claude-opus-4-5-thinking": {
      "quotaInfo": {
        "remainingFraction": 0.0,
        "resetTime": "2026-01-25T00:00:00Z"
      }
    }
  }
}

The 4 displayed models:

Display Name	Model Key	Alternate Key
G3 Pro	`gemini-3-pro-high`	`gemini-3-pro-low`
G3 Image	`gemini-3-pro-image`	-
G3 Flash	`gemini-3-flash`	-
Claude	`claude-opus-4-5-thinking`	`claude-opus-4-5`

Internal Data Types

Query Result Type

All platform query functions return a unified result format.

typescript

interface QueryResult {
  /** Whether successful */
  success: boolean;

  /** Output content on success */
  output?: string;

  /** Error message on failure */
  error?: string;
}

Constant Configuration

typescript

/** High usage warning threshold (percentage) */
export const HIGH_USAGE_THRESHOLD = 80;

/** API request timeout (milliseconds) */
export const REQUEST_TIMEOUT_MS = 10000;

Appendix: Source Code Reference

Click to expand source code locations

Last updated: 2026-01-23

Feature	File Path	Line
Auth data types	`plugin/lib/types.ts`	99-104
OpenAI auth	`plugin/lib/types.ts`	28-33
Zhipu AI auth	`plugin/lib/types.ts`	38-41
Copilot auth	`plugin/lib/types.ts`	46-51
Copilot PAT config	`plugin/lib/types.ts`	66-73
Antigravity accounts	`plugin/lib/types.ts`	78-94
OpenAI response format	`plugin/lib/openai.ts`	29-36
Zhipu AI response format	`plugin/lib/zhipu.ts`	43-50
Copilot internal API	`plugin/lib/copilot.ts`	47-58
Copilot Billing API	`plugin/lib/copilot.ts`	80-84
Google Cloud response	`plugin/lib/google.ts`	27-37

Key constants:

HIGH_USAGE_THRESHOLD = 80: High usage warning threshold (types.ts:111)
REQUEST_TIMEOUT_MS = 10000: API request timeout (types.ts:114)

Key types:

QueryResult: Query result type (types.ts:15-19)
CopilotTier: Copilot subscription type enum (types.ts:57)

快速开始

Platforms

Advanced

FAQ

附录

版本记录

Data Models: Auth File Structures & API Response Formats

What You'll Learn

Appendix Contents

Auth File Structures

OpenAI Auth Data

Zhipu AI / Z.ai Auth Data

GitHub Copilot Auth Data

Copilot PAT Configuration: `~/.config/opencode/copilot-quota-token.json`

Google Cloud Accounts: `~/.config/opencode/antigravity-accounts.json`

API Response Formats

OpenAI Response Format

Zhipu AI / Z.ai Response Format

GitHub Copilot Response Format

Method 1: Internal API (requires Copilot permission)

Method 2: Public Billing API (requires Fine-grained PAT)

Google Cloud Response Format

Internal Data Types

Query Result Type

Constant Configuration

Appendix: Source Code Reference

Data Models: Auth File Structures & API Response Formats ​

What You'll Learn ​

Appendix Contents ​

Auth File Structures ​

Main Auth File: ~/.local/share/opencode/auth.json ​

OpenAI Auth Data ​

Zhipu AI / Z.ai Auth Data ​

GitHub Copilot Auth Data ​

Copilot PAT Configuration: ~/.config/opencode/copilot-quota-token.json ​

Google Cloud Accounts: ~/.config/opencode/antigravity-accounts.json ​

API Response Formats ​

OpenAI Response Format ​

Zhipu AI / Z.ai Response Format ​

GitHub Copilot Response Format ​

Method 1: Internal API (requires Copilot permission) ​

Method 2: Public Billing API (requires Fine-grained PAT) ​

Google Cloud Response Format ​

Internal Data Types ​

Query Result Type ​

Constant Configuration ​

Appendix: Source Code Reference ​

Data Models: Auth File Structures & API Response Formats

What You'll Learn

Appendix Contents

Auth File Structures

Main Auth File: `~/.local/share/opencode/auth.json`

OpenAI Auth Data

Zhipu AI / Z.ai Auth Data

GitHub Copilot Auth Data

Copilot PAT Configuration: `~/.config/opencode/copilot-quota-token.json`

Google Cloud Accounts: `~/.config/opencode/antigravity-accounts.json`

API Response Formats

OpenAI Response Format

Zhipu AI / Z.ai Response Format

GitHub Copilot Response Format

Method 1: Internal API (requires Copilot permission)

Method 2: Public Billing API (requires Fine-grained PAT)

Google Cloud Response Format

Internal Data Types

Query Result Type

Constant Configuration

Appendix: Source Code Reference