Provider Configuration

This skill helps you configure LLM providers for Letta, supporting both:

• Self-hosted setup - Environment variables for Docker/local deployments
• BYOK via API - Programmatic provider management for Cloud and self-hosted

Quick Start

Add Provider via API

python scripts/setup_provider.py --type openai --api-key sk-...

Generate .env File for Docker

python scripts/generate_env.py --providers openai,anthropic,ollama

Validate Provider Credentials

python scripts/validate_provider.py --provider-id provider-xxx

Common Workflows

Workflow 1: Add OpenAI BYOK Provider

# Using the setup script
python scripts/setup_provider.py \
  --type openai \
  --name "My OpenAI" \
  --api-key sk-your-key-here

# Or via REST API directly
curl -X POST http://localhost:8283/v1/providers \
  -H "Content-Type: application/json" \
  -d '{
    "name": "My OpenAI",
    "provider_type": "openai",
    "api_key": "sk-your-key-here"
  }'

Workflow 2: Configure Self-Hosted Server

# Generate environment file
python scripts/generate_env.py --providers openai,anthropic,ollama --output .env

# Use with Docker
docker run -p 8283:8283 --env-file .env letta/letta:latest

Workflow 3: Validate and Refresh Provider

# Check credentials
python scripts/validate_provider.py --provider-id provider-abc123

# Refresh available models
curl -X PATCH http://localhost:8283/v1/providers/provider-abc123/refresh

Reference Documentation

Common Providers

For detailed OpenAI, Anthropic, Azure OpenAI, and Google AI configuration: Read references/common_providers.md

Self-Hosted Providers

For Ollama, vLLM, LM Studio, and other local providers: Read references/self_hosted_providers.md

All Provider Types

For comprehensive reference covering all 20+ supported providers: Read references/all_providers.md

Environment Variables

For Docker setup and multi-provider configuration: Read references/environment_variables.md

Scripts

setup_provider.py

Add providers programmatically via REST API.

Required flags:

• --type - Provider type (openai, anthropic, azure, etc.)
• --api-key - API key for the provider

Optional flags:

• --name - Display name (defaults to provider type)
• --base-url - Custom endpoint URL
• --api-version - API version (for Azure)
• --region - Region (for Bedrock)
• --access-key - Access key (for AWS)
• --server - Letta server URL (default: http://localhost:8283)

validate_provider.py

Check if provider credentials are valid and list available models.

Required flags:

• --provider-id - Provider ID to validate

Optional flags:

• --server - Letta server URL (default: http://localhost:8283)

generate_env.py

Generate .env file for Docker deployments.

Required flags:

• --providers - Comma-separated list of providers (openai,anthropic,ollama)

Optional flags:

• --output - Output file path (default: .env)
• --ollama-url - Ollama endpoint URL (default: http://host.docker.internal:11434)

Provider Types

Supported provider types:

• openai - OpenAI (GPT models)
• anthropic - Anthropic (Claude models)
• azure - Azure OpenAI
• google_ai - Google AI (Gemini)
• google_vertex - Google Vertex AI
• ollama - Ollama (local models)
• groq - Groq
• deepseek - DeepSeek
• xai - xAI (Grok)
• zai - Z.ai
• together - Together AI
• mistral - Mistral AI
• cerebras - Cerebras
• bedrock - AWS Bedrock
• vllm - vLLM
• sglang - SGLang
• hugging_face - Hugging Face
• lmstudio_openai - LM Studio (OpenAI compatible)
• minimax - MiniMax
• chatgpt_oauth - ChatGPT OAuth

For detailed configuration of each provider, see the reference files.

Notes

• All API-created providers are automatically BYOK (bring your own key)
• Provider IDs are prefixed with provider-
• After creating a provider, models are auto-discovered
• Use PATCH /v1/providers/{provider_id}/refresh to rescan models
• Environment variables are only for self-hosted deployments
• Letta Cloud uses BYOK providers exclusively (no env vars)