Build with Mintdesk

Quickstart

1. 1

   Create account

   Sign up with Google or email.

2. 2

   Add tokens

   Pick a prepaid pack and pay with UPI or cards.

3. 3

   Create a key

   Copy it once and set the base URL.

Setup guides

Three common ways to use Mintdesk: coding in the terminal, chatting in a chat app, and working with an AI agent inside your editor.

export ANTHROPIC_BASE_URL="https://api.mintdesk.tech"
export ANTHROPIC_AUTH_TOKEN="md_live_your_key"
claude

Anthropic-style apps:
  Base URL: https://api.mintdesk.tech
OpenAI-style apps:
  Base URL: https://api.mintdesk.tech/v1
API key:   md_live_your_key
Model:     claude-sonnet-5-5

Provider:  OpenAI-compatible
Base URL:  https://api.mintdesk.tech/v1
API key:   md_live_your_key
Model:     claude-opus-5-5

The Claude desktop and web apps do not offer a custom address setting, so they cannot connect to Mintdesk directly.

Authentication

Send your key as x-api-key or Authorization: Bearer. Keys start with md_live_ and are shown only once.

Base URL:  https://api.mintdesk.tech
OpenAI SDK / Cursor:  https://api.mintdesk.tech/v1
Models:    https://api.mintdesk.tech/v1/models
Header:    Authorization: Bearer md_live_your_key

Use https://api.mintdesk.tech when your client appends /v1/models. OpenAI SDKs and Cursor use https://api.mintdesk.tech/v1.

Endpoints

Code samples

VS CodeClaude Code CLICursorPythonNodecURL

export ANTHROPIC_BASE_URL="https://api.mintdesk.tech"
export ANTHROPIC_AUTH_TOKEN="md_live_your_key"
export ANTHROPIC_MODEL="claude-fable-5-1"

claude

Models

Streaming

Set "stream": true to receive server-sent events as the response is generated — the same event format Anthropic SDKs already understand.

curl https://api.mintdesk.tech/v1/messages \
  -H "x-api-key: md_live_your_key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-sonnet-5-5","max_tokens":512,"stream":true,
       "messages":[{"role":"user","content":"Hello"}]}'

No request limits

Mintdesk does not apply per-minute, hourly or daily request limits. Send requests as fast as your tools can — the only limit is your prepaid token balance.

Best practices

• • Keep keys on the server — never ship them in browser or mobile app code.
• • Create one key per project and give each its own token allocation.
• • Retry 429 and 5xx responses with exponential backoff.
• • Pick Haiku 5.5 for high-volume tasks and Opus 5.5 or Fable 5.1 Max for hard reasoning.

Billing & limits

• • Input and output tokens are counted against your shared prepaid balance; repeated context is billed at a lower cached rate.
• • Your balance never expires and works across every model.
• • Requests are only rejected with 402 when your balance reaches zero — top up to continue.

Error responses

Ready? Create your key →