Overview

Self-Hosting Documentation Access

This section requires a password to access. Interested in self-hosting? Contact sales to learn more.

Password:

Access Documentation

Self-Hosting Documentation Access Granted Logout

LlamaCloud supports multiple LLM models through different provider access methods to power its document parsing, extraction, and AI capabilities. This section provides guidance on configuring and choosing between different model providers for your self-hosted deployment.

Supported Models and Providers

Model Family	Developer Direct (Simple Setup)	Enterprise Cloud (Advanced Features)
OpenAI GPT GPT-4o, GPT-4.1, GPT-5, GPT-5	OpenAI API	Azure OpenAI
Anthropic Claude Claude 4.0 Sonnet, Claude 3.5 Haiku, Claude 3 Opus	Anthropic API	AWS Bedrock
Google Gemini Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.0 Flash	Google Gemini API	Google Vertex AI

Configuration Methods

Centralized Provider Configuration

Centralized provider configuration enables:

• Multiple provider instances with different credentials
• Custom API endpoints (e.g., for API gateways like Bifrost or Portkey)
• Custom HTTP headers per provider instance
• Custom provider-specific model names

Use config.llms.providerConfigs when the simple provider settings are not specific enough. It registers explicit model endpoints for gateways, custom endpoints, multiple credentials, custom headers, and feature-specific routing. Parsing tiers can also use it to control model order.

config:
  llms:
    providerConfigs:
      - id: "my-openai-config"
        provider: "openai"
        model_id: "openai-gpt-4o"
        provider_model_name: "@openai/gpt-4o"  # Optional: custom model name for gateway
        credentials:
          api_key: "sk-..."
          base_url: "https://api.gateway.com/v1"  # Optional: custom endpoint
        headers:  # Optional: custom headers
          x-api-key: "your-gateway-key"
      - id: "my-anthropic-config"
        provider: "anthropic"
        model_id: "anthropic-sonnet-4.5"
        provider_model_name: "@anthropic/claude-sonnet-4-5"  # Optional: custom model name for gateway
        credentials:
          api_key: "sk-ant-..."
          base_url: "https://api.gateway.com"  # Optional: custom endpoint
        headers:  # Optional: custom headers
          x-api-key: "your-gateway-key"

See: Centralized Provider Configuration Guide for the full schema, provider list, and parsing tier-ordering controls.

Provider-Specific Configuration

For basic direct-provider setup, use provider-specific configuration with Kubernetes secrets:

config:
  llms:
    openai:
      secret: <your-openai-secret>
    anthropic:
      secret: <your-anthropic-secret>
    gemini:
      secret: <your-gemini-secret>
    azureOpenAi:
      secret: <your-azureOpenAi-secret>
    awsBedrock:
      secret: <your-bedrock-secret>
    googleVertexAi:
      secret: <your-vertex-secret>

Helm Values Configuration (Legacy)

Some providers support direct configuration in Helm values (being deprecated):

backend:
  config:
    openAiApiKey: "your-api-key"

Next Steps

Configuration Guides

• Centralized Provider Configuration - For custom API gateways, multiple credentials, and parsing tier ordering

Provider-Specific Setup

Choose your LLM provider and follow the detailed setup instructions:

• OpenAI Setup
• Azure OpenAI Setup
• Anthropic API Setup
• AWS Bedrock Setup
• Google Gemini API Setup
• Google Vertex AI Setup

Troubleshooting

Verification Steps

After configuration, verify your setup by:

1. Using the LlamaCloud admin UI to confirm available models
2. Testing with a simple parsing or extraction task

Common Issues

1. Model not available: Check provider documentation for model availability in your region
2. Authentication failures: Verify API keys and permissions
3. Rate limiting: Monitor usage and implement appropriate quotas