# QuaLo Documentation

> Tổng hợp từ [qualo.xyz/docs](https://www.qualo.xyz/docs) — Cập nhật: April 2026

---

## Mục lục

- [Giới thiệu](#giới-thiệu)
- [Quickstart](#quickstart)
- [Xác thực (Authentication)](#xác-thực)
- [API Reference](#api-reference)
- [SDK](#sdk)
  - [OpenAI SDK](#openai-sdk)
  - [Anthropic SDK](#anthropic-sdk)
- [Tích hợp (Integrations)](#tích-hợp)
  - [Kilo Code](#kilo-code)
  - [Roo Code](#roo-code)
  - [Claude Code CLI](#claude-code-cli)
  - [Droid](#droid)
  - [OpenCode](#opencode)
  - [OpenClaw](#openclaw)
  - [Pi](#pi)

---

## Giới thiệu

Dịch vụ API xây dựng trên nền Anthropic, cung cấp quyền truy cập **tương thích OpenAI** tới Claude và các mô hình AI khác — thay thế trực tiếp, hỗ trợ streaming đầy đủ, không cần thay đổi SDK.

Bất kỳ client nào hoạt động với OpenAI đều sẽ hoạt động với QuaLo — chỉ cần thay đổi **base URL** và **API key**.

### Base URLs

| Định dạng | Base URL | Ghi chú |
|-----------|----------|---------|
| **OpenAI-compatible** | `https://llm.qualo.xyz/v1` | Có `/v1` |
| **Anthropic** | `https://llm.qualo.xyz` | **KHÔNG** có `/v1` |

### Models

| Model | Provider | Context | Output | Best For |
|-------|----------|---------|--------|----------|
| `claude-opus-4-6` | Anthropic | 200K | 64K | Complex reasoning, architecture |
| `claude-sonnet-4-6` | Anthropic | 200K | 64K | Balanced — recommended cho đa số tasks |
| `claude-haiku-4-5` | Anthropic | 144K | 64K | Speed-optimized, simple edits |
| `gpt-5.4` | OpenAI | 200K | 64K | Complex reasoning, vision |

### Ví dụ nhanh

```bash
curl https://llm.qualo.xyz/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-qualo-api-key" \
  -d '{
    "model": "gpt-5.4",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

---

## Quickstart

### 1. Lấy API Key

Đăng ký tại [qualo.xyz/register](https://qualo.xyz/register) → Dashboard → Copy API key (chỉ hiển thị một lần).

```bash
export QUALO_API_KEY=your-api-key-here
```

### 2. Gọi API đầu tiên

**OpenAI format:**

```bash
curl https://llm.qualo.xyz/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $QUALO_API_KEY" \
  -d '{
    "model": "gpt-5.4",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

**Anthropic format:**

```bash
curl https://llm.qualo.xyz/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: $QUALO_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

### 3. Streaming

**OpenAI SSE:**

```bash
curl https://llm.qualo.xyz/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $QUALO_API_KEY" \
  -d '{
    "model": "gpt-5.4",
    "stream": true,
    "messages": [{"role": "user", "content": "Tell me a short story."}]
  }'
```

**Anthropic SSE:**

```bash
curl https://llm.qualo.xyz/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: $QUALO_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 1024,
    "stream": true,
    "messages": [{"role": "user", "content": "Tell me a short story."}]
  }'
```

---

## Xác thực

### Authentication Headers

**OpenAI format (Bearer Token):**

```
Authorization: Bearer <your-api-key>
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36
```

**Anthropic format (x-api-key):**

```
x-api-key: <your-api-key>
anthropic-version: 2023-06-01
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36
```

### ⚠️ Bypass Cloudflare (Bắt buộc)

Tất cả requests **phải** bao gồm custom `User-Agent` header để tránh lỗi 403:

```
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36
```

### Biến môi trường

```bash
# OpenAI SDK compatibility
OPENAI_API_KEY=your-api-key-here
OPENAI_BASE_URL=https://llm.qualo.xyz/v1

# Anthropic SDK compatibility
ANTHROPIC_API_KEY=your-api-key-here
ANTHROPIC_BASE_URL=https://llm.qualo.xyz
```

### Kiểm tra Key

```bash
curl https://llm.qualo.xyz/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{"model": "gpt-5.4", "messages": [{"role": "user", "content": "ping"}], "max_tokens": 5}'
```

- ✅ HTTP 200 → Key hoạt động
- ❌ HTTP 401 → Key không hợp lệ hoặc thiếu

---

## API Reference

### Endpoints

| Method | Path | Mô tả |
|--------|------|--------|
| `POST` | `/v1/chat/completions` | OpenAI Chat Completions |
| `POST` | `/v1/messages` | Anthropic Messages API |
| `POST` | `/v1/responses` | OpenAI Responses API (o-series, Codex) |
| `GET` | `/v1/models` | Danh sách models |
| `POST` | `/v1/messages/count_tokens` | Đếm tokens |

### POST /v1/chat/completions

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `model` | string | ✅ | Model ID (e.g., `gpt-5.4`, `claude-sonnet-4-6`) |
| `messages` | array | ✅ | Array of `{role, content}` objects |
| `stream` | boolean | | SSE streaming. Default: `false` |
| `temperature` | number | | 0–2. Default: `1` |
| `max_tokens` | integer | | Max tokens to generate |
| `reasoning_effort` | string | | `low`, `medium`, `high` |

```bash
curl https://llm.qualo.xyz/v1/chat/completions \
  -H "Authorization: Bearer $QUALO_API_KEY" \
  -H "User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": false
  }'
```

**Response:**

```json
{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "created": 1714000000,
  "model": "gpt-5.4",
  "choices": [{
    "index": 0,
    "message": {"role": "assistant", "content": "Hello! How can I help you today?"},
    "finish_reason": "stop"
  }],
  "usage": {"prompt_tokens": 10, "completion_tokens": 12, "total_tokens": 22}
}
```

### POST /v1/messages

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `model` | string | ✅ | Anthropic model ID |
| `max_tokens` | integer | ✅ | Maximum tokens (bắt buộc) |
| `messages` | array | ✅ | Array of message objects |
| `system` | string/array | | System prompt |
| `stream` | boolean | | SSE streaming |
| `temperature` | number | | 0–1. Default: `1` |
| `thinking` | object | | Extended thinking: `{ type: "enabled", budget_tokens: N }` |

```bash
curl https://llm.qualo.xyz/v1/messages \
  -H "x-api-key: $QUALO_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

**Response:**

```json
{
  "id": "msg_abc123",
  "type": "message",
  "role": "assistant",
  "content": [{"type": "text", "text": "Hello! How can I help you today?"}],
  "model": "claude-sonnet-4-6",
  "stop_reason": "end_turn",
  "usage": {"input_tokens": 10, "output_tokens": 12}
}
```

### POST /v1/responses

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `model` | string | ✅ | Model ID (e.g., `o3`, `codex-mini-latest`) |
| `input` | string/array | ✅ | Prompt hoặc messages array |
| `instructions` | string | | System instructions |
| `stream` | boolean | | SSE streaming |
| `max_output_tokens` | integer | | Max output tokens |
| `reasoning` | object | | `{ effort: "low"|"medium"|"high", summary: "auto" }` |

### GET /v1/models

```bash
curl https://llm.qualo.xyz/v1/models \
  -H "Authorization: Bearer $QUALO_API_KEY" \
  -H "User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
```

### POST /v1/messages/count_tokens

```bash
curl https://llm.qualo.xyz/v1/messages/count_tokens \
  -H "x-api-key: $QUALO_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-4-6", "messages": [{"role": "user", "content": "Hello"}]}'
```

Response: `{ "input_tokens": 14 }`

### Error Codes

| HTTP Code | Meaning |
|-----------|---------|
| 400 | Bad request — JSON lỗi hoặc thiếu field |
| 401 | Unauthorized — API key sai hoặc thiếu |
| 403 | Forbidden — Cloudflare chặn (thiếu User-Agent) |
| 429 | Rate limit exceeded |
| 500 | Internal server error |
| 503 | Service unavailable — token pool cạn |

---

## SDK

### OpenAI SDK

```bash
pip install openai
```

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://llm.qualo.xyz/v1",
    api_key="your-qualo-api-key",
    default_headers={
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
    }
)

# Non-streaming
response = client.chat.completions.create(
    model="gpt-5.4",
    messages=[{"role": "user", "content": "Hello!"}],
    max_tokens=1024
)
print(response.choices[0].message.content)

# Streaming
stream = client.chat.completions.create(
    model="gpt-5.4",
    messages=[{"role": "user", "content": "Write a poem"}],
    stream=True
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

# List models
for model in client.models.list().data:
    print(model.id)
```

> ⚠️ Base URL **phải** bao gồm `/v1` khi dùng OpenAI SDK.

### Anthropic SDK

```bash
pip install anthropic
```

```python
from anthropic import Anthropic

client = Anthropic(
    base_url="https://llm.qualo.xyz",   # KHÔNG có /v1
    api_key="your-qualo-api-key",
    default_headers={
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
    }
)

# Non-streaming
response = client.messages.create(
    model="claude-sonnet-4-6",
    messages=[{"role": "user", "content": "Hello!"}],
    max_tokens=1024
)
print(response.content[0].text)

# Streaming
with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=256,
    messages=[{"role": "user", "content": "Write a poem"}]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

# System message
response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system="You are a helpful coding assistant.",
    messages=[{"role": "user", "content": "How do I read a file in Python?"}]
)
```

> ⚠️ Anthropic SDK base URL **KHÔNG** có `/v1`. SDK tự thêm `/v1/messages`.

---

## Tích hợp

### Kilo Code

**VS Code Extension**

#### Dùng Claude (Anthropic):

1. Mở VS Code → Settings → Extensions → Kilo Code
2. Cấu hình:

```json
{
  "kilocode.apiProvider": "anthropic",
  "kilocode.apiBaseUrl": "https://llm.qualo.xyz",
  "kilocode.apiKey": "your-qualo-api-key",
  "kilocode.model": "claude-sonnet-4-6"
}
```

#### Dùng GPT/Gemini (OpenAI-compatible):

```json
{
  "kilocode.apiProvider": "openai-compatible",
  "kilocode.apiBaseUrl": "https://llm.qualo.xyz/v1",
  "kilocode.apiKey": "your-qualo-api-key",
  "kilocode.model": "gpt-5.4"
}
```

#### Troubleshooting

- **Connection refused:** Kiểm tra base URL — Anthropic: không có `/v1`, OpenAI-compatible: có `/v1`
- **Invalid API key:** Verify key đúng và còn active
- **Model not found:** Dùng đúng tên: `claude-opus-4-6`, `claude-sonnet-4-6`, `claude-haiku-4-5`, `gpt-5.4`

---

### Roo Code

**VS Code Extension**

#### Dùng Claude (Anthropic):

```json
{
  "rooCode.apiProvider": "anthropic",
  "rooCode.anthropic.baseUrl": "https://llm.qualo.xyz",
  "rooCode.anthropic.apiKey": "your-qualo-api-key",
  "rooCode.anthropic.model": "claude-sonnet-4-6"
}
```

#### Dùng GPT/Gemini (OpenAI-compatible):

```json
{
  "rooCode.apiProvider": "openai-compatible",
  "rooCode.openaiCompatible.baseUrl": "https://llm.qualo.xyz/v1",
  "rooCode.openaiCompatible.apiKey": "your-qualo-api-key",
  "rooCode.openaiCompatible.model": "gpt-5.4"
}
```

#### Troubleshooting

- **No completions:** Verify base URL kết thúc `/v1` (OpenAI) hoặc không (Anthropic), API key đúng
- **Slow responses:** Dùng `claude-haiku-4-5` cho inline completions nhanh hơn

---

### Claude Code CLI

**Terminal AI assistant từ Anthropic**

#### 1. Cài đặt

```bash
npm install -g @anthropic-ai/claude-code
```

#### 2. Cấu hình `~/.claude/settings.json`

```json
{
  "env": {
    "ANTHROPIC_BASE_URL": "https://llm.qualo.xyz",
    "ANTHROPIC_AUTH_TOKEN": "your-qualo-api-key",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-4-6",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-4-6",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-4-5",
    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1"
  },
  "permissions": {
    "allow": [],
    "deny": []
  },
  "alwaysThinkingEnabled": true
}
```

#### 3. Chạy

```bash
cd your-project
claude
# Kiểm tra: gõ /status
```

#### Model Variables

| Variable | Default |
|----------|---------|
| `ANTHROPIC_DEFAULT_OPUS_MODEL` | `claude-opus-4-6` |
| `ANTHROPIC_DEFAULT_SONNET_MODEL` | `claude-sonnet-4-6` |
| `ANTHROPIC_DEFAULT_HAIKU_MODEL` | `claude-haiku-4-5` |

#### Troubleshooting

- **Auth error:** Kiểm tra `ANTHROPIC_AUTH_TOKEN`. Thử xóa `~/.claude/settings.json` và config lại.
- **Timeout:** Tăng `API_TIMEOUT_MS` lên `600000` hoặc cao hơn
- **Connection error:** Kiểm tra internet, đảm bảo `https://llm.qualo.xyz` không bị firewall chặn

---

### Droid

**AI Software Engineering Agent by Factory**

#### Environment Variables

```bash
export ANTHROPIC_BASE_URL="https://llm.qualo.xyz"
export ANTHROPIC_API_KEY="your-qualo-api-key"
```

#### Config file `~/.droid/config.json`

```json
{
  "provider": "anthropic",
  "model": "claude-sonnet-4-6",
  "apiBaseUrl": "https://llm.qualo.xyz",
  "maxTokens": 8192,
  "temperature": 0.7
}
```

#### Alternative: Custom Models

```json
{
  "custom_models": [
    {
      "model_display_name": "Claude Sonnet 4.6",
      "model": "claude-sonnet-4-6",
      "base_url": "https://llm.qualo.xyz",
      "api_key": "your-qualo-api-key",
      "provider": "anthropic"
    }
  ]
}
```

#### Test

```bash
droid
# hoặc
droid "What's 2+2?"
```

#### Troubleshooting

- **Connection error:** Base URL là `https://llm.qualo.xyz` (không có `/v1`)
- **Model not responding:** Kiểm tra API key còn credits
- **Slow responses:** Dùng `claude-haiku-4-5` hoặc giảm `maxTokens`

---

### OpenCode

**TUI-rich Terminal AI Coding Agent**

#### Config file: `opencode.json` (project root hoặc `~/.config/opencode/opencode.json`)

#### Dùng GPT/Gemini (OpenAI-compatible):

```json
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "qualo": {
      "name": "QuaLo (OpenAI)",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "apiKey": "your-qualo-api-key",
        "baseURL": "https://llm.qualo.xyz/v1"
      }
    }
  }
}
```

#### Dùng Claude (Anthropic):

```json
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "qualo-anthropic": {
      "name": "QuaLo (Anthropic)",
      "npm": "@ai-sdk/anthropic",
      "options": {
        "apiKey": "your-qualo-api-key",
        "baseURL": "https://llm.qualo.xyz/v1"
      }
    }
  }
}
```

#### Full config với models:

```json
{
  "provider": {
    "qualo": {
      "name": "QuaLo (OpenAI)",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "apiKey": "your-qualo-api-key",
        "baseURL": "https://llm.qualo.xyz/v1"
      },
      "models": {
        "gpt-5.4": {
          "name": "GPT-5.4",
          "limit": { "context": 200000, "output": 64000 },
          "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] },
          "reasoning": true,
          "options": { "reasoningEffort": "xhigh" },
          "tool_call": true,
          "attachment": true
        }
      }
    },
    "qualo-anthropic": {
      "name": "QuaLo (Anthropic)",
      "npm": "@ai-sdk/anthropic",
      "options": {
        "apiKey": "your-qualo-api-key",
        "baseURL": "https://llm.qualo.xyz/v1"
      },
      "models": {
        "claude-opus-4-6": {
          "name": "Claude Opus 4.6",
          "limit": { "context": 200000, "output": 64000 },
          "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] },
          "reasoning": true,
          "tool_call": true,
          "attachment": true
        },
        "claude-sonnet-4-6": {
          "name": "Claude Sonnet 4.6",
          "limit": { "context": 200000, "output": 32000 },
          "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] },
          "reasoning": true,
          "tool_call": true,
          "attachment": true
        },
        "claude-haiku-4-5": {
          "name": "Claude Haiku 4.5",
          "limit": { "context": 144000, "output": 32000 },
          "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] },
          "reasoning": true,
          "tool_call": true
        }
      }
    }
  }
}
```

#### Chọn default model:

```json
{ "model": "qualo-anthropic/claude-sonnet-4-6" }
// hoặc
{ "model": "qualo/gpt-5.4" }
```

#### Troubleshooting

- **Model not appearing:** Kiểm tra model ID đúng trong provider's models block
- **Auth error:** Kiểm tra apiKey khớp với key từ QuaLo dashboard
- **Wrong base URL:** GPT/Gemini → `/v1`, Anthropic/Claude → `/v1` (OpenCode dùng `/v1` cho cả hai)

---

### OpenClaw

**Multi-model Agentic Coding**

#### Dùng Claude (Anthropic):

```json
{
  "provider": "anthropic",
  "baseUrl": "https://llm.qualo.xyz",
  "apiKey": "your-qualo-api-key",
  "model": "claude-sonnet-4-6"
}
```

#### Dùng GPT/Gemini (OpenAI-compatible):

```json
{
  "provider": "openai-compatible",
  "baseUrl": "https://llm.qualo.xyz/v1",
  "apiKey": "your-qualo-api-key",
  "model": "gpt-5.4"
}
```

#### Multi-provider config `~/.openclaw/config.json`:

```json
{
  "providers": {
    "qualo": {
      "type": "openai-compatible",
      "baseUrl": "https://llm.qualo.xyz/v1",
      "apiKey": "your-qualo-api-key"
    },
    "qualo-anthropic": {
      "type": "anthropic",
      "baseUrl": "https://llm.qualo.xyz",
      "apiKey": "your-qualo-api-key"
    }
  },
  "defaultProvider": "qualo-anthropic",
  "defaultModel": "claude-sonnet-4-6"
}
```

> OpenClaw gửi requests trực tiếp — không cần thêm User-Agent header.

#### Troubleshooting

- **Connection refused:** OpenAI-compatible → `/v1`, Anthropic → không có `/v1`
- **Model not found:** Dùng đúng ID: `claude-opus-4-6`, `claude-sonnet-4-6`, `claude-haiku-4-5`, `gpt-5.4`

---

### Pi

**Terminal AI Coding Agent (`@mariozechner/pi-coding-agent`)**

#### 1. Tạo `~/.pi/agent/models.json`

```bash
mkdir -p ~/.pi/agent
```

#### 2. Nội dung `models.json`:

```json
{
  "providers": {
    "qualo": {
      "baseUrl": "https://llm.qualo.xyz/v1",
      "api": "openai-completions",
      "apiKey": "your-qualo-api-key",
      "models": [
        {
          "id": "gpt-5.4",
          "name": "GPT-5.4",
          "reasoning": true,
          "input": ["text", "image"],
          "contextWindow": 200000,
          "maxTokens": 64000
        }
      ]
    },
    "qualo-anthropic": {
      "baseUrl": "https://llm.qualo.xyz",
      "api": "anthropic-messages",
      "apiKey": "your-qualo-api-key",
      "models": [
        {
          "id": "claude-opus-4-6",
          "name": "Claude Opus 4.6",
          "reasoning": true,
          "input": ["text", "image"],
          "contextWindow": 144000,
          "maxTokens": 64000
        },
        {
          "id": "claude-sonnet-4-6",
          "name": "Claude Sonnet 4.6",
          "reasoning": true,
          "input": ["text", "image"],
          "contextWindow": 200000,
          "maxTokens": 32000
        },
        {
          "id": "claude-haiku-4-5",
          "name": "Claude Haiku 4.5",
          "reasoning": true,
          "input": ["text", "image"],
          "contextWindow": 144000,
          "maxTokens": 32000
        }
      ]
    }
  }
}
```

#### API Key formats:

```json
"apiKey": "sk-qualo-..."           // literal value
"apiKey": "QUALO_API_KEY"          // env var name
"apiKey": "!echo $QUALO_API_KEY"   // shell command
```

#### 3. Chọn model

```bash
pi
```

- `Ctrl+L` — mở model picker
- `Ctrl+P` — cycle qua favourite models
- `/model` — tương đương Ctrl+L

> `models.json` reload mỗi lần mở `/model` — không cần restart.

#### Troubleshooting

- **Models không hiện trong /model:** Kiểm tra `~/.pi/agent/models.json` là valid JSON, `baseUrl` đúng
- **Auth error:** Verify API key — có thể dùng literal, env var name, hoặc shell command

---

## Tham khảo nhanh

| Tool | Provider | Base URL | Config Key |
|------|----------|----------|------------|
| **Kilo Code** | `anthropic` | `https://llm.qualo.xyz` | `kilocode.apiBaseUrl` |
| **Kilo Code** | `openai-compatible` | `https://llm.qualo.xyz/v1` | `kilocode.apiBaseUrl` |
| **Roo Code** | `anthropic` | `https://llm.qualo.xyz` | `rooCode.anthropic.baseUrl` |
| **Roo Code** | `openai-compatible` | `https://llm.qualo.xyz/v1` | `rooCode.openaiCompatible.baseUrl` |
| **Claude Code** | anthropic | `https://llm.qualo.xyz` | `ANTHROPIC_BASE_URL` |
| **Droid** | anthropic | `https://llm.qualo.xyz` | `ANTHROPIC_BASE_URL` |
| **OpenCode** | `@ai-sdk/openai-compatible` | `https://llm.qualo.xyz/v1` | `options.baseURL` |
| **OpenCode** | `@ai-sdk/anthropic` | `https://llm.qualo.xyz/v1` | `options.baseURL` |
| **OpenClaw** | `anthropic` | `https://llm.qualo.xyz` | `baseUrl` |
| **OpenClaw** | `openai-compatible` | `https://llm.qualo.xyz/v1` | `baseUrl` |
| **Pi** | `anthropic-messages` | `https://llm.qualo.xyz` | `baseUrl` |
| **Pi** | `openai-completions` | `https://llm.qualo.xyz/v1` | `baseUrl` |

> **Quy tắc chung:** OpenAI-compatible → có `/v1` | Anthropic → **KHÔNG** có `/v1` (trừ OpenCode dùng `/v1` cho cả hai)