AiStore AI API

Quickstart

1. Create an account, add money and buy AI credits.
2. Open API Keys and generate your secret key (sk-…).
3. Make your first call:

curl https://mail.aishop.bd/api/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6-luna",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Authentication

Send your key as a bearer token on every request. One active key per account — regenerating immediately revokes the previous key.

Authorization: Bearer sk-YOUR_KEY

Base URLs

Two compatible surfaces, one key. Pick whichever your client speaks.

Chat completions (OpenAI format)

POST https://mail.aishop.bd/api/v1/chat/completions

Supports streaming ("stream": true). Both GPT-family and Claude-family models work through this one endpoint.

curl https://mail.aishop.bd/api/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6-luna",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Explain SSE in one sentence."}
    ],
    "stream": false
  }'

Response

{
  "id": "chatcmpl-…",
  "object": "chat.completion",
  "model": "gpt-6-luna",
  "choices": [
    {
      "index": 0,
      "message": {"role": "assistant", "content": "Server-sent events stream…"},
      "finish_reason": "stop"
    }
  ],
  "usage": {"prompt_tokens": 21, "completion_tokens": 14, "total_tokens": 35},
  "x_cost_usd": 0.0012
}

When streaming, you receive chat.completion.chunk objects followed by a final chunk carrying usage and x_cost_usd, then a data: [DONE] sentinel.

Messages (Anthropic format)

POST https://mail.aishop.bd/api/v1/messages

curl https://mail.aishop.bd/api/v1/messages \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 1024,
    "system": "You are a helpful assistant.",
    "messages": [
      {"role": "user", "content": "Hello, Claude!"}
    ]
  }'

Response

{
  "id": "msg_…",
  "type": "message",
  "role": "assistant",
  "model": "claude-sonnet-4-6",
  "content": [{"type": "text", "text": "Hello! How can I help?"}],
  "stop_reason": "end_turn",
  "usage": {"input_tokens": 12, "output_tokens": 9}
}

With "stream": true you get the standard Anthropic event sequence: message_start → content_block_delta → message_stop.

Use with Claude Code

Run Claude Code — the terminal coding agent — on your AiStore credits. No Anthropic subscription or foreign card needed. Tool calls, streaming and token usage are fully supported on every Claude model.

export ANTHROPIC_BASE_URL=https://mail.aishop.bd/api
export ANTHROPIC_AUTH_TOKEN=sk-YOUR_KEY
export ANTHROPIC_MODEL=claude-sonnet-4-6
claude

$env:ANTHROPIC_BASE_URL = "https://mail.aishop.bd/api"
$env:ANTHROPIC_AUTH_TOKEN = "sk-YOUR_KEY"
$env:ANTHROPIC_MODEL = "claude-sonnet-4-6"
claude

Cline / VS Code setup

Coding agents that live inside VS Code — Cline, Kilo Code, Roo Code — run on your AiStore credits too. Pick the Anthropic provider and point its base URL at this API: tool calls, file edits and streaming work exactly like the official providers, with no foreign card.

Base URL: https://mail.aishop.bd/api/v1
API key:  sk-YOUR_KEY
Model:    any chat model id from the list below

Models

GET https://mail.aishop.bd/api/v1/models

Image generation (async job)

POST https://mail.aishop.bd/api/v1/images/generations

Image, audio, and video generation run as async jobs: submit a task, then poll until status is success or failed. Jobs are billed upfront and refunded automatically on failure.

curl https://mail.aishop.bd/api/v1/images/generations \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ideogram/v3-text-to-image",
    "prompt": "a cozy cabin in snowy woods at dusk",
    "aspect_ratio": "16:9",
    "n": 1
  }'

# → 202 Accepted
{ "id": 42, "object": "generation.job", "status": "processing", "price": 0.05, ... }

Common params: model, prompt, aspect_ratio, n (1–4). seedream/5-pro-text-to-image prices by quality tier — basic (1K) $0.05 / high (2K) $0.09 per image — and also accepts output_format (png|jpeg) and nsfw_checker (boolean). Models may cap n — Seedream generates a single image per request. seedream/5-flash-layer-decomposition splits a supplied image_url (required) into independent layers — it takes size (auto|1K|1.5K|2K) instead of aspect_ratio and bills $0.02 per returned image (every layer, including the base image, counts as one): one image is held upfront and the difference settled when the job completes.

Audio generation (async job)

POST https://mail.aishop.bd/api/v1/audio/generations

curl https://mail.aishop.bd/api/v1/audio/generations \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs/ten-v3-text-to-speech",
    "prompt": "Welcome to our store!",
    "voice": "optional-voice-id",
    "format": "mp3"
  }'

Voice-tuned speech (ElevenLabs voice_tuning models — billed per 1,000 characters)

curl https://mail.aishop.bd/api/v1/audio/generations \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs/text-to-speech-turbo-2-5",
    "prompt": "Unlock powerful API with Kie.ai!",
    "voice": "EkK5I93UQWFDigLMpZcX",
    "stability": 0.5,
    "similarity_boost": 0.75,
    "style": 0,
    "speed": 1,
    "language_code": "en",
    "nsfw_checker": true
  }'

• elevenlabs/text-to-speech-turbo-2-5 (≈$0.04/1,000 chars) and elevenlabs/text-to-speech-multilingual-v2 (≈$0.08/1,000 chars) accept the voice-tuning params above, plus optional timestamps (boolean) and previous_text / next_text for chunk continuity. 422 on models without voice tuning.
• stability, similarity_boost, style are 0–1; speed is 0.7–1.2; language_code picks the language on multilingual models (e.g. en, bn).

Multi-speaker dialogue (dialogue TTS models)

curl https://mail.aishop.bd/api/v1/audio/generations \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-3-8-flash-lite-tts",
    "speakers": [
      { "speaker_id": "Speaker 1", "voice_name": "Fola" },
      { "speaker_id": "Speaker 2", "voice_name": "Alnilam" }
    ],
    "dialogue_turns": [
      { "speaker_id": "Speaker 1", "text": "Halt, traveler! The northern pass is sealed." },
      { "speaker_id": "Speaker 2", "style": "whisper quietly", "text": "Then I will take the river path." }
    ],
    "filler_words": true,
    "temperature": 1
  }'

• dialogue_turns — 1–50 ordered turns (speaker_id, optional style up to 500 chars, text up to 2,000); every speaker_id must appear in speakers (max 2, unique ids — or omit speakers and every turn speaker gets the model's default voice). Dialogue fields are only accepted by models that list them — a plain prompt becomes a single-speaker turn there.
• filler_words (boolean) / temperature (0–2) — optional delivery controls for dialogue models.
• Send either prompt or dialogue_turns, never both.

Video generation (async job)

POST https://mail.aishop.bd/api/v1/videos/generations

curl https://mail.aishop.bd/api/v1/videos/generations \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan/2-7-text-to-video",
    "prompt": "a drone shot over misty mountains at sunrise",
    "duration": 5,
    "resolution": "1080p",
    "aspect_ratio": "16:9"
  }'

• resolution / duration accepted values come from the model's row in the table above (duration 5–15s, up to 30s on models that list it).
• image_urls — optional array (max 2) of public image URLs for image-to-video / reference models (e.g. grok-imagine/image-to-video, pixverse-v6/reference-to-video); required there, ignored by text-to-video models.
• audio — optional boolean, models with the audio flag only (e.g. wan/3-0-video-prime): generates music + dialogue baked into the video. 422 on models without it.

Polling a job

GET https://mail.aishop.bd/api/v1/generations/{id}

curl https://mail.aishop.bd/api/v1/generations/42 \
  -H "Authorization: Bearer sk-YOUR_KEY"

# → while running
{ "id": 42, "status": "processing", "urls": [], ... }

# → when finished (URLs valid ~24 hours)
{ "id": 42, "status": "success", "price": 0.05,
  "urls": ["https://…/image.png"], ... }

# → on failure (charge refunded automatically)
{ "id": 42, "status": "failed", "error": "…", "price": 0.05, "urls": [] }

Store & wallet (run a shop from any bot or app)

Browse the catalog

GET https://mail.aishop.bd/api/v1/store/products

Every active product on the website with its pricing options and live stock — everything a bot needs to render a shop menu.

curl https://mail.aishop.bd/api/v1/store/products \
  -H "Authorization: Bearer sk-YOUR_KEY"

# →
{ "object": "list", "data": [
  { "object": "store.product", "id": 7, "slug": "chatgpt-plus",
    "title": "ChatGPT Plus", "brand": "OpenAI",
    "description": "…", "image": "https://…/chatgpt.jpg",
    "delivery_type": "manual",
    "stock": { "type": "private_account", "available": 12, "in_stock": true },
    "prices": [
      { "id": 21, "duration": "1 Month", "price": 19.0, "is_default": true },
      { "id": 22, "duration": "3 Months", "price": 49.0, "is_default": false }
    ],
    "url": "https://mail.aishop.bd/products/chatgpt-plus" } ] }

One product: GET https://mail.aishop.bd/api/v1/store/products/{slug} — same shape, plus the product_price_id values the purchase call below expects. Upstream (api-type) products report available: null.

Check the wallet

GET https://mail.aishop.bd/api/v1/wallet

curl https://mail.aishop.bd/api/v1/wallet \
  -H "Authorization: Bearer sk-YOUR_KEY"

# → { "object": "wallet", "balance": 42.5, "currency": "USD", "ai_credits": 1000 }

balance is the main USD wallet used for store purchases; ai_credits is what AI requests deduct.

Buy a product

POST https://mail.aishop.bd/api/v1/store/products/{slug}/purchase

Buy the products shown on the website (subscriptions, accounts, license keys) with your wallet balance — the same instant-delivery flow as the storefront.

curl https://mail.aishop.bd/api/v1/store/products/chatgpt-plus/purchase \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "product_price_id": 12,
    "quantity": 1,
    "coupon_code": "SAVE5"
  }'

Response

{
  "object": "store.order",
  "order_no": "ASN-4TZ9WQK2M",
  "status": "completed",
  "product": "ChatGPT Plus",
  "duration": "1 Month",
  "quantity": 1,
  "subtotal": 19.0,
  "discount": 5.0,
  "total": 14.0,
  "currency": "USD",
  "wallet_balance": 36.0,
  "delivery_token": "dt-…",
  "delivery_url": "https://mail.aishop.bd/api/v1/store/delivery/dt-…",
  "delivery_expires_at": "2026-10-06T12:00:00+00:00"
}

Credentials are never part of the purchase response. Reveal them with a second call:

GET https://mail.aishop.bd/api/v1/store/delivery/{delivery_token}

curl https://mail.aishop.bd/api/v1/store/delivery/dt-… \
  -H "Authorization: Bearer sk-YOUR_KEY"

# → the purchased credentials
{ "object": "store.delivery", "order_no": "ASN-…", "status": "completed",
  "credentials": [
    { "product": "ChatGPT Plus", "type": "account",
      "email": "user@example.com", "password": "…" }
  ] }

Order history

GET https://mail.aishop.bd/api/v1/store/orders

curl https://mail.aishop.bd/api/v1/store/orders \
  -H "Authorization: Bearer sk-YOUR_KEY"

# →
{ "object": "list", "data": [
  { "order_no": "ASN-4TZ9WQK2M", "status": "completed", "product": "ChatGPT Plus",
    "duration": "1 Month", "quantity": 1, "total": 14.0, "currency": "USD",
    "created_at": "2026-10-05T12:00:00+00:00",
    "detail_url": "https://mail.aishop.bd/api/v1/store/orders/ASN-4TZ9WQK2M" } ] }

GET https://mail.aishop.bd/api/v1/store/orders/{order_no} returns one order with its full credentials array — the same reveal as the delivery endpoint, but with no expiry (the My Orders equivalent, always available to the key owner).

• The delivery token is valid for 24 hours and only works with your own API key. After it expires (410), the credentials stay available in My Orders on the website — and via the order-history endpoint above.
• Coupon codes from the storefront work here too (case-insensitive).
• Failed orders never debit the wallet — stock shortages roll back automatically.
• Rate limits: catalog/wallet/orders reads 60 per minute, purchases 10 per minute.

Pricing & billing

• Credits: every AI request — chat, images, video, audio — deducts AI credits (list USD price × the credits-per-USD rate shown on the credit packages page). x_cost_usd fields stay in USD.
• Chat: metered per token — input and output prices are per 1M tokens (see the table above). Credits are deducted after each response using the upstream-reported usage. Requests your credits clearly can't cover are rejected with 402 before any call is made.
• Images: flat price × number of images, charged upfront.
• Video: per-second price (by resolution) × duration, charged upfront.
• Audio: flat price per generation, charged upfront — except dialogue TTS models (e.g. google/gemini-3-8-flash-lite-tts), which are priced from the script length (≈ the model row's price per 1,000 characters, with a small per-generation minimum).
• Refunds: if a job fails, the credits are returned automatically (ledger type refund).
• Buy credits anytime from your wallet page (paid from the main wallet, topped up via the deposit page); every debit appears in the ledger as ai_usage.

Errors

{ "error": { "message": "Insufficient AI credits…", "type": "ai_gateway_error" } }

Rate limits

• Chat endpoints: 60 requests / minute
• Image & audio generation: 20 requests / minute
• Video generation: 10 requests / minute

Cline / VS Code setup

Streaming works in both — tokens arrive live as they are generated.