LLM

Multi-provider LLM client with cascade fallback. The LLM class provides a unified interface for multiple LLM providers with automatic fallback on failures.

---

Classes

  members:
    - chat
    - chat_json
  show_root_heading: true
  show_source: true

---

  show_root_heading: true
  show_source: true

---

Usage

Auto-Detect Providers

```python from web3_agent_kit.llm import LLM

Automatically detects providers from environment variables

llm = LLM()

Cascade order: Anthropic → Kimi → OpenRouter → DeepSeek → Groq → OpenAI

response = llm.chat("What is the best yield on Base?") print(response) ```

JSON Response

```python data = llm.chat_json("Analyze this swap: 0.1 ETH to USDC on Base") print(data)

{

"analysis": "Good swap, low slippage expected",

"estimated_output": 350.0,

"gas_estimate": "0.002 ETH"

}

```

Custom Configuration

python from web3_agent_kit.llm import LLM, LLMConfig config = LLMConfig( providers=[ { "name": "anthropic", "api_key": "sk-ant-...", "model": "claude-sonnet-4-20250514", "base_url": "https://api.anthropic.com/v1", }, { "name": "openai", "api_key": "sk-...", "model": "gpt-4o-mini", "base_url": "https://api.openai.com/v1", }, ], temperature=0.1, max_tokens=2048, timeout=30, ) llm = LLM(config=config)

---

Supported Providers

Provider	Env Variable	Default Model
Anthropic	ANTHROPIC_API_KEY	claude-sonnet-4-20250514
OpenAI	OPENAI_API_KEY	gpt-4o-mini
Groq	GROQ_API_KEY	llama-3.3-70b-versatile
DeepSeek	DEEPSEEK_API_KEY	deepseek-chat
OpenRouter	OPENROUTER_API_KEY	anthropic/claude-sonnet-4
Kimi	KIMI_API_KEY	moonshot-v1-8k
---
## Error Handling
The LLM client automatically cascades through providers on failure:
- 429 (Rate Limited) — Moves to next provider
- 5xx (Server Error) — Moves to next provider
- Timeout — Moves to next provider
If all providers fail, raises RuntimeError.