AI Framework

Model Providers Configuration

Configure OpenAI, Anthropic, Google Gemini, Groq, NVIDIA NIM, Ollama, and local models in Vista.

Updated: 2026-03-20

Vista AI includes built-in drivers for the most popular language model providers, using standard HTTP fetch under the hood.

Provider Specifiers

Specify models using a simple provider:model-name format:

typescript

typescript
// OpenAI
model: 'openai:gpt-4o';
model: 'openai:gpt-4o-mini';

// Anthropic
model: 'anthropic:claude-3-5-sonnet-20241022';
model: 'anthropic:claude-3-5-haiku-20241022';

// Google Gemini
model: 'gemini:gemini-1.5-pro';
model: 'gemini:gemini-1.5-flash';

// Local / Ollama
model: 'ollama:llama3';
model: 'ollama:mistral';

// Groq
model: 'groq:llama-3.1-8b-instant';

// NVIDIA NIM
model: 'nvidia:meta/llama-3.1-8b-instruct';

// Mock / Testing
model: 'mock:echo';

Environment Variables

Configure API keys and base URLs in your .env.local or deployment platform settings:

.env.local

bash
# OpenAI
OPENAI_API_KEY="sk-..."
OPENAI_BASE_URL="https://api.openai.com/v1" # optional

# Anthropic
ANTHROPIC_API_KEY="sk-ant-..."
ANTHROPIC_BASE_URL="https://api.anthropic.com/v1" # optional

# Google Gemini
GEMINI_API_KEY="AIza..."

# Ollama / Local
OLLAMA_BASE_URL="http://localhost:11434/v1" # optional (default)

# Groq
GROQ_API_KEY="gsk_..."

# NVIDIA NIM
NVIDIA_API_KEY="nvapi-..."

Embeddings

ts

ts
import { embedText, createRetrieverTool, InMemoryVectorStore } from 'vista/ai';

const store = new InMemoryVectorStore();
store.addDocument({
  id: 'intro',
  text: 'Vista is a React framework.',
  vector: await embedText('Vista is a React framework.', { model: 'openai:text-embedding-3-small' }),
});

export const search = createRetrieverTool({
  store,
  embed: (query) => embedText(query, { model: 'openai:text-embedding-3-small' }),
});

Custom Provider Options

You can also pass custom configuration directly to resolveModel or your agent:

typescript

typescript
import { agent, createOpenAIModel } from '@vistagenic/vista/ai';

export const customAgent = agent({
  name: 'custom',
  model: createOpenAIModel({
    model: 'deepseek-chat',
    baseURL: 'https://api.deepseek.com/v1',
    apiKey: process.env.DEEPSEEK_API_KEY,
  }),
});

Related