93 lines
2.2 KiB
YAML
93 lines
2.2 KiB
YAML
# LLM Provider Registry
|
|
# AI/LLM provider configurations for Advisory AI
|
|
|
|
version: "1.0"
|
|
category: llm-providers
|
|
|
|
# Global LLM settings
|
|
defaults:
|
|
enabled: false # Explicitly enable providers
|
|
timeout: "00:02:00"
|
|
maxRetries: 3
|
|
|
|
# ============================================================================
|
|
# LLM PROVIDERS
|
|
# ============================================================================
|
|
providers:
|
|
# Cloud providers
|
|
claude:
|
|
enabled: false
|
|
priority: 100
|
|
config: claude.yaml
|
|
description: "Anthropic Claude (Claude 3.5/4)"
|
|
mode: remote
|
|
|
|
openai:
|
|
enabled: false
|
|
priority: 90
|
|
config: openai.yaml
|
|
description: "OpenAI GPT-4/4o"
|
|
mode: remote
|
|
|
|
azure-openai:
|
|
enabled: false
|
|
priority: 90
|
|
config: azure-openai.yaml
|
|
description: "Azure OpenAI Service"
|
|
mode: remote
|
|
|
|
# Local providers (for air-gap)
|
|
ollama:
|
|
enabled: false
|
|
priority: 80
|
|
config: ollama.yaml
|
|
description: "Ollama local inference"
|
|
mode: local
|
|
|
|
llama-server:
|
|
enabled: false
|
|
priority: 80
|
|
config: llama-server.yaml
|
|
description: "llama.cpp HTTP server"
|
|
mode: local
|
|
|
|
vllm:
|
|
enabled: false
|
|
priority: 80
|
|
config: vllm.yaml
|
|
description: "vLLM inference server"
|
|
mode: local
|
|
|
|
# ============================================================================
|
|
# INFERENCE SETTINGS
|
|
# ============================================================================
|
|
inference:
|
|
# Mode: remote, local, hybrid
|
|
mode: "${ADVISORY_AI_INFERENCE_MODE:-local}"
|
|
|
|
# Fallback chain
|
|
fallbackChain:
|
|
- claude
|
|
- openai
|
|
- ollama
|
|
|
|
# Model selection
|
|
modelSelection:
|
|
# Task-specific model overrides
|
|
explanation: "" # Use default
|
|
remediation: "" # Use default
|
|
classification: "" # Use default
|
|
|
|
# ============================================================================
|
|
# OFFLINE/AIR-GAP SETTINGS
|
|
# ============================================================================
|
|
offline:
|
|
# Signed model bundle path
|
|
modelBundlePath: "${ADVISORY_AI_MODEL_BUNDLE_PATH:-/opt/stellaops/offline/models}"
|
|
|
|
# Verify bundle signatures
|
|
verifySignatures: true
|
|
|
|
# Public key for signature verification
|
|
publicKeyPath: "${ADVISORY_AI_MODEL_PUBKEY:-/etc/stellaops/model-signing-pubkey.pem}"
|