@agentix-e/nl2spel-openai

OpenAI-compatible LLM Provider for @agentix-e/nl2spel

7 preset providers · Custom API support · Streaming output

API Documentation

→ nl2spel-openai API

Preset Providers (7)

| Provider | Model | Context | Streaming | |----------|-------|---------|-----------| | OpenAI | gpt-4o-mini | 128K | ✅ | | DeepSeek | deepseek-chat | 64K | ✅ | | GLM (Zhipu) | glm-4-flash | 128K | ✅ | | Copilot | gpt-4o | 128K | ✅ | | Hunyuan (Tencent) | hunyuan-lite | 32K | ✅ | | MiniMax | abab6.5s-chat | 32K | ✅ | | Kimi | moonshot-v1-8k | 8K | ✅ |

Quick Start

import { NL2SpelEngine } from '@agentix-e/nl2spel';
import { OpenAICompatibleProvider } from '@agentix-e/nl2spel-openai';

const engine = new NL2SpelEngine();

// OpenAI
engine.registerProvider(new OpenAICompatibleProvider({
  provider: 'openai',
  apiKey: 'sk-...',
}));

// DeepSeek (best value)
engine.registerProvider(new OpenAICompatibleProvider({
  provider: 'deepseek',
  apiKey: 'sk-...',
}));

// Custom API
engine.registerProvider(new OpenAICompatibleProvider({
  custom: {
    name: 'my-llm',
    baseURL: 'https://my-llm.example.com/v1',
    apiKey: 'sk-...',
    model: 'my-model-v1',
  },
}));

Multi-Provider Fallback

engine.registerProvider(new OpenAICompatibleProvider({ provider: 'openai', apiKey: '...' }));
engine.registerProvider(new OpenAICompatibleProvider({ provider: 'deepseek', apiKey: '...' }));

// Auto fallback by cost from low to high
const result = await engine.generate('Complex multi-level nested expression');

License

MIT © 2025 Agentix-E