AI Gateway

Singdata AI Gateway is an enterprise-grade AI gateway service that provides unified multi-model API management, intelligent routing and scheduling, BYOK model integration, and usage analytics—helping enterprises simplify multi-vendor LLM integration with a single connection that can call all models.

Access Methods

AI Gateway supports two access scenarios:

Pre-configured clients (zero configuration, ready out of the box): The following clients have AI Gateway integration built in with no additional configuration required:

Client	Purpose
Lakehouse SQL	Call AI_COMPLETE() / AI_EMBEDDING() directly in SQL
Data Analytics Agent	Natural language conversational data analysis
Data Engineering Agent	Natural language ETL development, task management, and ops diagnostics
cz-cli	CLI command line / MCP tool interface
Singclaw	Desktop intelligent agent that understands your business

Using Lakehouse SQL as an example, simply call AI_COMPLETE() or AI_EMBEDDING() directly in SQL to use all enabled models:

-- Text generation, sentiment analysis, translation, summarization... SELECT AI_COMPLETE('gw:qwen-max', CONCAT('Summarize in one sentence: ', content)) FROM articles; -- Vectorization for semantic search and RAG SELECT AI_EMBEDDING('gw:text-embedding-v3', text) FROM documents;

Custom Agent (one Endpoint + one API Key): Use the standard OpenAI SDK, replace base_url and api_key, no other code changes needed:

from openai import OpenAI client = OpenAI( base_url="https://<your-instance>.singdata.com/ai-gateway/v1", api_key="<your-api-key>", ) response = client.chat.completions.create( model="moonshotai/kimi-k2.6", messages=[{"role": "user", "content": "hello"}], )

What I Want to Do

Goal	Where to Go
Understand AI Gateway features and positioning	Product Introduction
Complete setup and call the first model in 10 minutes	Quick Start
View supported models and pricing	Model Pricing
Integrate your own third-party model (BYOK)	Product Introduction → BYOK
Manage API Keys, set usage limits	Product Introduction → API Key Management
Call LLMs in SQL for data analysis	Call LLMs in SQL
Analyze images in SQL	Analyze Images in SQL

Core Capabilities at a Glance

Unified access: One API endpoint to call models from 20+ providers including Qwen, DeepSeek, GLM, Kimi, Doubao, GPT, Claude, Gemini, etc. 100% OpenAI interface compatible; existing code only needs endpoint and key replaced.

Intelligent routing: Automatically selects the optimal model by price, latency, and throughput, with support for automatic failover and load balancing.

BYOK: Bring your own third-party API Key; billing goes directly to the vendor account with no additional platform charges.

Usage control: Track token consumption by API Key, team, and project dimensions, with support for limits and alerts, and real-time cost breakdown viewing.

Quick Start

→ Complete initial setup and call your first model in 10 minutes