Build with leading AI models through one API key.

Quickstart

Use this base URL everywhere:

https://api.ezkielyna.store/v1

Send your first request (replace sk-YOUR_KEY with your key):

curl https://api.ezkielyna.store/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-v4.1-flash","stream":true,"messages":[{"role":"user","content":"Hello!"}]}'

API keys & groups

Every key belongs to a group. The group decides which models the key can use and the price multiplier. You pick it when you create the key.

OpenAI format

Point any OpenAI SDK or tool at https://api.ezkielyna.store/v1. Supported endpoints: /chat/completions, /embeddings, /rerank, /images/generations, /audio/speech, /audio/transcriptions, /videos and /models.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.ezkielyna.store/v1",
    api_key="sk-YOUR_KEY",
)

stream = client.chat.completions.create(
    model="deepseek-v4.1-flash",
    messages=[{"role": "user", "content": "Explain RAG in two sentences."}],
    stream=True,
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.ezkielyna.store/v1",
  apiKey: "sk-YOUR_KEY",
});

const stream = await client.chat.completions.create({
  model: "deepseek-v4.1-flash",
  messages: [{ role: "user", content: "Explain RAG in two sentences." }],
  stream: true,
});
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}

curl https://api.ezkielyna.store/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "stream": true,
    "messages": [{"role": "user", "content": "Explain RAG in two sentences."}]
  }'

List the models your key can use: GET https://api.ezkielyna.store/v1/models.

Anthropic format

Every chat model also speaks the Anthropic Messages API. Use base URL https://api.ezkielyna.store (the SDK adds /v1/messages) and send your key as x-api-key.

from anthropic import Anthropic

client = Anthropic(
    base_url="https://api.ezkielyna.store",
    api_key="sk-YOUR_KEY",
)

message = client.messages.create(
    model="claude-sonnet-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a haiku about Jakarta."}],
)
print(message.content[0].text)

curl https://api.ezkielyna.store/v1/messages \
  -H "x-api-key: sk-YOUR_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-5",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Write a haiku about Jakarta."}]
  }'

Any model works here — for example deepseek-v4-pro or glm-5.3 through the Anthropic format as well.

Streaming & reasoning tips

• Use "stream": true. You see the first words sooner, and long answers never time out. Non-stream requests also work, but must finish within 100 seconds (otherwise you get 524), so always stream long or reasoning-heavy answers.
• Reasoning models think first. Models such as kimi-k3, glm-5.3, hy4 or deepseek-v4-pro spend tokens on a thinking phase. Give them enough max_tokens (2,000+), or the visible answer can be empty.
• Go faster and cheaper with "reasoning_effort": "low" for simple tasks (chat, summaries, translations).
• Big prompts take longer. A 100k-token prompt usually needs 6–12 seconds before the first token. Keep context short when you can.
• Prompt caching is automatic on supported models and makes repeated context up to 4× cheaper.
• Images: base64 data: URLs work on every vision model; plain https:// image URLs can fail on some models. Requests can be up to 32 MB.
• Base URL: https://api.ezkielyna.store/v1 for OpenAI-style clients, https://api.ezkielyna.store for Anthropic/Claude Code. Not supported: /v1/messages/count_tokens and legacy /v1/completions.
• Need speed? Try glm-5.3-flash or deepseek-v4.1-flash (first token in ~1–2 s).

Embeddings, images, video & audio

Embeddings & rerank (RAG)

curl https://api.ezkielyna.store/v1/embeddings \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"bge-m3","input":["Halo dunia","Hello world"]}'

curl https://api.ezkielyna.store/v1/rerank \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"bge-reranker-base","query":"capital of Indonesia",
       "documents":["Jakarta is the capital of Indonesia.","Cats like fish."],"top_n":2}'

Good picks: bge-m3 or qwen3-embedding-0.6b (multilingual, great for Indonesian) and bge-reranker-base for reranking.

Images

curl https://api.ezkielyna.store/v1/images/generations \
  -H "Authorization: Bearer sk-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-image-2","prompt":"a cozy coffee shop in Bandung, watercolor"}'

Image and video models are in the Media group: use a Media or Auto key. A Standard or Premium key cannot call them.

You get a 1024×1024 image link that stays online for 24 hours — download it right away, or add "response_format":"b64_json" to receive the image data directly. Models include gpt-image-2, nano-banana-2, flux.2-max, ideogram-4 and grok-imagine-image-pro; the full list is in Model Square under the Media group.

Video

Video is generated in the background. Submit a job, poll its status, then download the MP4. Use a Media or Auto key.

# 1) submit (seconds: 4, 8 or 12 · size: 1280x720 or 720x1280)
curl https://api.ezkielyna.store/v1/videos \
  -H "Authorization: Bearer sk-YOUR_KEY" -H "Content-Type: application/json" \
  -d '{"model":"grok-imagine-video","prompt":"a red kite over the sea","seconds":"4","size":"1280x720"}'

# 2) check status until "completed"
curl https://api.ezkielyna.store/v1/videos/TASK_ID -H "Authorization: Bearer sk-YOUR_KEY"

# 3) download
curl -L -o video.mp4 https://api.ezkielyna.store/v1/videos/TASK_ID/content -H "Authorization: Bearer sk-YOUR_KEY"

Video is billed per second of footage.

Text-to-speech

curl https://api.ezkielyna.store/v1/audio/speech \
  -H "Authorization: Bearer sk-YOUR_KEY" -H "Content-Type: application/json" \
  -d '{"model":"edge-tts","input":"Halo, selamat datang!","voice":"id-ID-GadisNeural"}' \
  -o speech.mp3

edge-tts is a Standard model. Voice: an Edge voice ID such as id-ID-GadisNeural, id-ID-ArdiNeural or en-US-AriaNeural, or an OpenAI voice name such as alloy. Output is MP3, up to 4,096 characters per request, billed per character.

One-click apps

The fastest way to chat: open the Chat menu in the console sidebar, pick an app and choose one of your keys. The app opens with Ezkielyna already configured.

Ezkielyna Studio

studio.ezkielyna.store is a browser workspace for your Ezkielyna API key: chat with any text model, create images and videos, and keep prompts, canvases and assets in one place. Charges come from your normal API balance.

1. Create a key in API Keys with the Auto group (it includes Media, where the image and video models live). Trial and promo accounts cannot create Media keys.
2. Open Studio → Account & settings → paste the key → Connect your account. The key is kept server-side for the session only, never in the browser.
3. Check the estimated charge in the confirm dialog, then generate. Results stay in your history for 7 days — download what you want to keep.

Studio is available in English, Chinese and Bahasa Indonesia (/id/). Step-by-step guide in Indonesian: Panduan → Ezkielyna Studio.

Claude Code, Codex & IDEs

Claude Code

Easiest: use CC Switch from the Chat menu. Or set these environment variables before running claude:

export ANTHROPIC_BASE_URL="https://api.ezkielyna.store"
export ANTHROPIC_AUTH_TOKEN="sk-YOUR_KEY"
export ANTHROPIC_MODEL="claude-sonnet-5"
export ANTHROPIC_DEFAULT_SONNET_MODEL="claude-sonnet-5"
export ANTHROPIC_DEFAULT_OPUS_MODEL="claude-opus-5.5"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="deepseek-v4.1-flash"
claude

$env:ANTHROPIC_BASE_URL = "https://api.ezkielyna.store"
$env:ANTHROPIC_AUTH_TOKEN = "sk-YOUR_KEY"
$env:ANTHROPIC_MODEL = "claude-sonnet-5"
$env:ANTHROPIC_DEFAULT_SONNET_MODEL = "claude-sonnet-5"
$env:ANTHROPIC_DEFAULT_OPUS_MODEL = "claude-opus-5.5"
$env:ANTHROPIC_DEFAULT_HAIKU_MODEL = "deepseek-v4.1-flash"
claude

Want a cheaper coding model? Set ANTHROPIC_MODEL to kimi-k3, glm-5.3 or deepseek-v4-pro — they all work through the Anthropic format.

Codex CLI

Add a provider to ~/.codex/config.toml and set your key as an environment variable:

model = "deepseek-v4.1-flash"
model_provider = "ezkielyna"

[model_providers.ezkielyna]
name = "Ezkielyna API"
base_url = "https://api.ezkielyna.store/v1"
env_key = "EZKIELYNA_API_KEY"
wire_api = "responses"

export EZKIELYNA_API_KEY="sk-YOUR_KEY"
codex

Cursor

1. Open Settings → Models.
2. Under OpenAI API Key, paste your key and enable Override OpenAI Base URL → https://api.ezkielyna.store/v1.
3. Click Add model and type the exact model name, e.g. claude-sonnet-5 or deepseek-v4-pro.

Cline, Roo Code & Kilo Code (VS Code)

1. API Provider: OpenAI Compatible.
2. Base URL: https://api.ezkielyna.store/v1 · API Key: your key.
3. Model ID: any model name from Model Square, e.g. kimi-k3.

OpenCode

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ezkielyna": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Ezkielyna API",
      "options": { "baseURL": "https://api.ezkielyna.store/v1", "apiKey": "sk-YOUR_KEY" },
      "models": { "kimi-k3": {}, "glm-5.3": {}, "claude-sonnet-5": {} }
    }
  }
}

Continue

models:
  - name: Ezkielyna DeepSeek
    provider: openai
    model: deepseek-v4-pro
    apiBase: https://api.ezkielyna.store/v1
    apiKey: sk-YOUR_KEY

Other apps

Any app with an “OpenAI-compatible” option works with the same three values:

This includes Open WebUI (Settings → Connections → OpenAI API), NextChat, n8n (OpenAI credentials → Base URL), Dify, Flowise, LangChain and LlamaIndex.

Models & pricing

Model Square lists every model with its live price. You pay only for what you use:

• Chat — input tokens + output tokens (cached input is cheaper).
• Images — per image. Video — per second. Transcription — per file. Text-to-speech — per character.
• The group rate is applied on top (Standard 1×, Premium 1.5×, Media 1×, Hemat 1×, Pro 3×).

What $10 gets you

Rough estimates with a 4:1 input:output ratio and no caching.

Balance, top up & refunds

• Top up: buy an Ezkielyna API Credit pack at ezkielyna.store (QRIS, Binance Pay, crypto or bot balance). You receive a redeem code — enter it in Wallet.
• Track usage: every request (model, tokens, cost, speed) appears in Usage Logs. Your balance is on the Dashboard.
• Failed requests: if a normal-size streamed request fails before any output is produced, the charge is refunded automatically within about 10 minutes.
• Promo bonus: sign-up bonus balance works on Standard models. When you top up, the same amount of bonus unlocks for Premium models too. Trial and promo accounts cannot create Media keys, so image and video need a top-up first.

Troubleshooting

FAQ & support

Still stuck? Email [email protected], message the admin on WhatsApp wa.me/6285710999204, ask in the Telegram community t.me/EzkielynaCommunity or use the bot @EzkielynaStore_Bot. Live model health: status page.