Other Providers

​

z.ai

​

Features

• Tokens limit (primary window)
• Time limit / MCP window (secondary, if tokens also present)
• API region selection (global or BigModel China mainland)
• Custom quota URL override

​

Authentication

export Z_AI_API_KEY="your-token-here"

{
  "providers": [
    {
      "id": "zai",
      "apiKey": "your-token-here"
    }
  ]
}

​

API Endpoints

• Global: GET https://api.z.ai/api/monitor/usage/quota/limit
• China: GET https://open.bigmodel.cn/api/monitor/usage/quota/limit
• Custom: Set Z_AI_QUOTA_URL environment variable
• Host override: Set Z_AI_API_HOST environment variable

​

Kimi

​

Features

• Weekly request quota tracking
• 5-hour rate limit (300 minutes)
• Membership tier display (Andante, Moderato, Allegretto)
• Automatic and manual authentication

​

Authentication

export KIMI_AUTH_TOKEN="jwt-token-here"

​

API Details

• Endpoint: POST https://www.kimi.com/apiv2/kimi.gateway.billing.v1.BillingService/GetUsages
• Authentication: Bearer token from kimi-auth cookie

​

Membership Tiers

​

Kilo

​

Features

• Auto-fallback from API to CLI auth
• Plan and activity tracking
• Configurable source mode (auto, api, cli)

​

Authentication

export KILO_API_KEY="your-api-key"

{
  "providers": [
    {
      "id": "kilo",
      "apiKey": "your-api-key"
    }
  ]
}

​

Source Modes

• Auto (default): Tries API first, falls back to CLI on 401/403 or missing credentials
• API: API token only
• CLI: CLI session only

​

API Details

• Endpoint: https://app.kilo.ai/api/trpc
• Authentication: Bearer token from config or environment

​

Kimi K2

​

Features

• Credit usage tracking (consumed / remaining)
• No explicit reset timestamp

​

Authentication

export KIMI_K2_API_KEY="your-api-key"
# or
export KIMI_API_KEY="your-api-key"
# or
export KIMI_KEY="your-api-key"

​

API Details

• Endpoint: GET https://kimi-k2.ai/api/user/credits
• Headers: Authorization: Bearer <api_key>, Accept: application/json
• Response header: X-Credits-Remaining

​

Kiro

​

Features

• Monthly credits percentage tracking
• Bonus credits tracking
• Plan name display (e.g., “KIRO FREE”)
• AWS Builder ID authentication via CLI

​

Authentication

1. Install kiro-cli
2. Log in with AWS Builder ID:
   kiro-cli auth login
   
3. CodexBar runs kiro-cli chat --no-interactive "/usage"

​

CLI Command

kiro-cli chat --no-interactive "/usage"

​

Output Format

• Monthly credits: percentage bar + reset date
• Bonus credits: consumed/total + expiry days

​

Status Page

• Manual link to AWS Health Dashboard
• No auto-polling

​

Vertex AI

​

Features

• Cloud Monitoring quota usage (time-series metrics)
• Token cost tracking from local Claude logs (Vertex AI entries)
• Automatic project discovery

​

Authentication

1. Install and configure gcloud:
   gcloud auth application-default login
   gcloud config set project PROJECT_ID
   
2. CodexBar reads application_default_credentials.json from gcloud config directory

​

API Endpoints

• Cloud Monitoring timeSeries:
  • Usage: serviceruntime.googleapis.com/quota/allocation/usage
  • Limit: serviceruntime.googleapis.com/quota/limit
  • Resource: consumer_quota with service="aiplatform.googleapis.com"

​

Token Cost Tracking

1. Model name format (primary): Vertex AI uses @ as version separator
   • Vertex AI: claude-opus-4-5@20251101
   • Anthropic: claude-opus-4-5-20251101
2. Metadata fields (fallback): metadata.provider: "vertexai" or keys containing vertex/gcp

1. Enable Vertex AI provider in Settings → Providers
2. Enable “Show cost summary” in Settings → General
3. Use Claude Code with Vertex AI model names (with @ format)

​

Environment Variables

• GOOGLE_CLOUD_PROJECT
• GCLOUD_PROJECT
• CLOUDSDK_CORE_PROJECT

​

Augment

​

Features

• Credits tracking and monthly limits
• Usage monitoring (credits consumed)
• Plan information display
• CLI integration (auggie account status)
• Automatic session keepalive (prevents cookie expiration)
• Multi-browser support: Chrome, Chrome Beta, Chrome Canary, Arc, Safari

​

Authentication

​

Cookie Names

• _session (legacy)
• auth0, auth0.is.authenticated, a0.spajs.txs (Auth0)
• __Secure-next-auth.session-token, next-auth.session-token (NextAuth)
• __Host-authjs.csrf-token, authjs.session-token (AuthJS)
• session, web_rpc_proxy_session (Augment-specific)

​

Session Keepalive

• Check interval: Every 5 minutes
• Refresh buffer: Refreshes 5 minutes before cookie expiration
• Rate limiting: Minimum 2 minutes between refresh attempts
• Session cookies: Refreshed every 30 minutes (no expiration date)

​

API Endpoints

• Credits: https://app.augmentcode.com/api/credits
• Subscription: https://app.augmentcode.com/api/subscription

​

Amp

​

Features

• Amp Free daily usage meter
• Time-to-full reset countdown
• Hourly replenishment rate

​

Authentication

​

API Details

• Endpoint: https://ampcode.com/settings
• Parses embedded freeTierUsage payload from HTML
• Computes time-to-full from hourly replenishment rate

​

JetBrains AI

​

Features

• Monthly credits tracking
• Refill date countdown
• Auto-detects installed JetBrains IDEs
• Uses most recently modified IDE

​

Supported IDEs

• IntelliJ IDEA
• PyCharm
• WebStorm
• GoLand
• CLion
• DataGrip
• RubyMine
• Rider
• PhpStorm
• RustRover
• Android Studio
• Fleet
• Aqua
• DataSpell

​

Data Source

• macOS: ~/Library/Application Support/JetBrains/
• macOS (Android Studio): ~/Library/Application Support/Google/
• Linux: ~/.config/JetBrains/
• Linux (Android Studio): ~/.config/Google/

​

XML Structure

• quotaInfo attribute (JSON): type, current, maximum, tariffQuota.available, until
• nextRefill attribute (JSON): type, next, tariff.amount, tariff.duration

​

Settings

• IDE Picker: Auto-detected IDEs list or “Auto-detect” (default)
• Custom Path: Manual base path override

​

Constraints

• Requires JetBrains IDE with AI Assistant enabled
• XML file only exists after AI Assistant usage
• Internal file format may change between IDE versions

​

OpenRouter

​

Features

• Credit usage percentage tracking
• Balance display (total credits - usage)
• Rate limit information
• Multi-provider AI model access (OpenAI, Anthropic, Google, Meta, etc.)

​

Authentication

export OPENROUTER_API_KEY="sk-or-v1-..."

{
  "providerConfig": {
    "openrouter": {
      "apiKey": "sk-or-v1-..."
    }
  }
}

​

API Endpoints

1. Credits API (/api/v1/credits):
   • Returns total credits purchased and total usage
   • Balance = total_credits - total_usage
2. Key API (/api/v1/key):
   • Returns rate limit information

​

Environment Variables

​

Notes

• Credit values are cached on OpenRouter’s side and may be up to 60 seconds stale
• Pre-purchase credits for usage
• Rate limits: 10+ credits = 1000 free model requests/day

​

Status Page

• Manual link to status.openrouter.ai
• No auto-polling yet

​

Antigravity

​

Features

• Model-specific quota tracking (Claude, Gemini Pro Low, Gemini Flash)
• Local-only (no external authentication)
• Automatic port discovery
• HTTPS and HTTP support

​

How It Works

1. Process detection:
   • Finds language_server_macos process with Antigravity markers
   • Extracts --csrf_token and --extension_server_port flags
2. Port discovery:
   • Runs lsof -nP -iTCP -sTCP:LISTEN -p <pid>
   • Probes all listening ports
3. Connect port probe (HTTPS):
   • POST https://127.0.0.1:<port>/exa.language_server_pb.LanguageServerService/GetUnleashData
   • Headers: X-Codeium-Csrf-Token: <token>, Connect-Protocol-Version: 1
   • First 200 OK response selects the connect port
4. Quota fetch:
   • Primary: GetUserStatus
   • Fallback: GetCommandModelConfigs
   • Retries over HTTP on extension_server_port if HTTPS fails

​

Model Mapping Priority

1. Claude (label contains claude but not thinking)
2. Gemini Pro Low (label contains pro + low)
3. Gemini Flash (label contains gemini + flash)
4. Fallback: lowest remaining percent

​

UI Mapping

• Primary: Claude quota
• Secondary: Gemini Pro quota
• Tertiary: Gemini Flash quota

​

Constraints

• Internal protocol (fields may change)
• Requires lsof for port detection
• Local HTTPS uses self-signed cert (insecure TLS allowed)
• Requires Antigravity app running

​

Status Page

• Google Workspace incidents for Gemini product

​

Droid (Factory)

​

Features

• Standard and Premium token usage tracking
• Billing window with reset date
• Multiple authentication fallback paths
• WorkOS token minting support

​

Authentication Methods (Fallback Order)

1. Cached cookie header (Keychain)
2. Stored session (~/Library/Application Support/CodexBar/factory-session.json)
3. Stored bearer token (same session file)
4. Stored WorkOS refresh token (same session file)
5. Local storage WorkOS tokens (Safari + Chrome/Chromium/Arc leveldb)
6. Browser cookies (Safari only) for Factory domains
7. WorkOS cookies (Safari) to mint tokens
8. Browser cookies (Chrome, Firefox) for Factory domains
9. WorkOS cookies (Chrome, Firefox) to mint tokens

​

Cookie Domains

• factory.ai
• app.factory.ai
• auth.factory.ai

​

Session Cookie Names

• wos-session
• __Secure-next-auth.session-token
• next-auth.session-token
• __Secure-authjs.session-token
• __Host-authjs.csrf-token
• authjs.session-token
• session
• access-token

​

API Endpoints

• https://auth.factory.ai
• https://api.factory.ai
• https://app.factory.ai

1. GET <baseURL>/api/app/auth/me
   • Org + subscription metadata + feature flags
2. POST <baseURL>/api/organization/subscription/usage
   • Body: { "useCache": true, "userId": "<id?>" }
   • Standard + Premium token usage and billing window

​

WorkOS Token Minting

POST https://api.workos.com/user_management/authenticate

• client_id: one of client_01HXRMBQ9BJ3E7QSTQ9X2PHVB7 or client_01HNM792M5G5G1A2THWPXKFMXB
• grant_type: refresh_token
• refresh_token: from local storage or session store
• Optional: organization_id, useCookie: true

​

Local Storage WorkOS Tokens

• Root: ~/Library/Containers/com.apple.Safari/Data/Library/WebKit/WebsiteData/Default
• Finds origin files containing app.factory.ai or auth.factory.ai
• Reads LocalStorage/localstorage.sqlite3

• Roots: ~/Library/Application Support/<Browser>/User Data/<Profile>/Local Storage/leveldb
• Scans LevelDB files for workos:refresh-token and workos:access-token

​

Snapshot Mapping

• Primary: Standard usage ratio
• Secondary: Premium usage ratio
• Reset: Billing period end date

​

Status Page

• status.factory.ai

​

OpenCode

​

Features

• Rolling 5-hour usage window (session)
• Weekly usage window
• Automatic workspace detection
• Browser cookie import with caching

​

Authentication

export CODEXBAR_OPENCODE_WORKSPACE_ID="wrk_..."
# or full URL
export CODEXBAR_OPENCODE_WORKSPACE_ID="https://opencode.ai/workspace/..."

​

API Endpoints

• POST https://opencode.ai/_server
  • Workspaces: function ID def39973159c7f0483d8793a822b8dbb10d067e12c65455fcb4608459ba0234f
  • Subscription: function ID 7abeebee372f304e050aaaf92be863f4a86490e382f8c79db68fd94040d691b4

​

Usage Mapping

• Primary window: Rolling 5-hour usage (rollingUsage.usagePercent, resets in rollingUsage.resetInSec)
• Secondary window: Weekly usage (weeklyUsage.usagePercent, resets in weeklyUsage.resetInSec)

​

Notes

• Responses are text/javascript with serialized objects (parsed via regex)
• Cookie import defaults to Chrome-only to avoid browser prompts
• Cached in Keychain (com.steipete.codexbar.cache, account cookie.opencode)

​

Warp

​

Features

• Monthly credits usage tracking
• Next refresh time display
• Unlimited plan detection
• Token-based authentication

​

Authentication

export WARP_API_KEY="wk-..."
# or
export WARP_TOKEN="wk-..."

​

API Endpoint

• POST https://app.warp.dev/graphql/v2?op=GetRequestLimitInfo
• Query: GetRequestLimitInfo
• Fields: isUnlimited, nextRefreshTime, requestLimit, requestsUsedSinceLastRefresh

​

Usage Mapping

• Credits: requestsUsedSinceLastRefresh / requestLimit
• Reset: nextRefreshTime
• Unlimited plans: Shows “Unlimited” with full bar

​

Troubleshooting

​

Reference

• Warp API Keys Documentation

​

Ollama

​

Features

• Plan tier detection (Free/Pro/Max)
• Session usage window
• Weekly usage window
• Automatic reset time tracking

​

Authentication

​

How It Works

• Scrapes https://ollama.com/settings using browser cookies
• Parses:
  • Plan badge from Cloud Usage header
  • Session and weekly usage percentages
  • Reset timestamps from data-time attributes

​

Usage Mapping

• Session: Percentage from session usage bar
• Weekly: Percentage from weekly usage bar
• Resets: ISO timestamps from data-time attributes

​

Troubleshooting

​

MiniMax

​

Features

• Coding plan usage (available vs. total)
• Multiple authentication methods with fallback
• Region selection (Global or China mainland)
• Browser local storage token extraction

​

Authentication

export MINIMAX_API_KEY="your-token"

​

API Endpoints

• API token: GET https://api.minimax.io/v1/coding_plan/remains
  • Requires Authorization: Bearer <api_token>
• Global web: https://platform.minimax.io
• China web: https://platform.minimaxi.com
• Coding plan page: {host}/user-center/payment/coding-plan
• Remains API: GET {host}/v1/api/openplatform/coding_plan/remains

​

Fallback Order

1. API token (if provided)
2. Cached cookie header (Keychain)
3. Browser cookie import (Safari → Chrome → Firefox)
4. Browser local storage access token (Chromium)
5. Manual cookie header override

​

Environment Overrides

• MINIMAX_HOST - Set to platform.minimaxi.com for China
• MINIMAX_CODING_PLAN_URL - Full URL override for coding plan page
• MINIMAX_REMAINS_URL - Full URL override for remains API
• MINIMAX_COOKIE or MINIMAX_COOKIE_HEADER - Manual cookie

​

Usage Mapping

• Primary: Percent used from model_remains (used/total)
• Window: Derived from start_time/end_time
• Reset: From remains_time or end_time

​

Notes

• Cookies alone often fail (status 1004); remains API expects Bearer token
• MiniMax stores access_token in Chromium local storage (LevelDB)
• JWT tokens are automatically extracted when using browser auth

​

Next Steps