ODS V3 Pre-Release: Public testing and refinement ahead of the official V3 launch. Turn your PC, Mac, or Linux box into a private AI server.
Python ai-agentsamdcomfyuidockerllama-cppllmlocal-ain8nnvidiaopen-webuiragself-hostedspeech-to-textstrix-halotext-to-speechworkflow-automation
Backend Runtime Contract
ODS now defines backend runtime behavior in contract files instead of hardcoded installer branches.
Contract Files
config/backends/amd.jsonconfig/backends/nvidia.jsonconfig/backends/cpu.jsonconfig/backends/apple.json
Each contract defines:
- LLM engine/service name
- public API port and health URL
- provider name + internal OpenAI-compatible provider URL (kept for compatibility; no installer phase reads them)
Loader
scripts/load-backend-contract.sh
Example:
source lib/safe-env.sh
BACKEND_ENV="$(scripts/load-backend-contract.sh --backend amd --env)"
load_env_from_output <<< "$BACKEND_ENV"
echo "$BACKEND_PUBLIC_HEALTH_URL $BACKEND_PROVIDER_URL"
Installer Integration
The modular installer loads backend contracts in installers/lib/detection.sh via load_backend_contract(). Contract values drive:
- runtime health-check endpoint selection (
installers/phases/12-health.sh) - LLM API summary endpoint (
installers/phases/13-summary.sh)
See INSTALLER-ARCHITECTURE.md for the full module map.