docker-compose.ollama-local.yml

Rama release/v3.8.54 · réplica de lectura del repositorio original.

# ──────────────────────────────────────────────────────────────────────
#  OmniRoute — Local Ollama (Docker → host-native Ollama)
# ──────────────────────────────────────────────────────────────────────
#
#  Use when Ollama runs natively on the Linux host and OmniRoute runs in
#  Docker. Requires host.docker.internal → host-gateway (Linux Docker 20.10+).
#
#  Usage:
#    docker compose -f docker-compose.ollama-local.yml up -d
#
#  Then in VS Code (OmniCopilot) or any OpenAI-compatible client:
#    base URL: http://<host-ip>:20128
#    model:    ollama/<your-model>
#
#  Configure the Local Ollama openai-compatible provider in the dashboard
#  with base URL: http://host.docker.internal:11434/v1
#
#  Timeout notes:
#    - v3.8.51+ auto-raises direct response-start timeout to 300s for local
#      targets (host.docker.internal, 192.168.x.x, localhost).
#    - Override via OMNIROUTE_LOCAL_DIRECT_HEADERS_TIMEOUT_MS if needed.
# ──────────────────────────────────────────────────────────────────────

services:
  omniroute:
    image: diegosouzapw/omniroute:latest
    container_name: omniroute
    environment:
      - OMNIROUTE_DIRECT_HEADERS_TIMEOUT_MS=${OMNIROUTE_DIRECT_HEADERS_TIMEOUT_MS:-300000}
      - OMNIROUTE_LOCAL_DIRECT_HEADERS_TIMEOUT_MS=${OMNIROUTE_LOCAL_DIRECT_HEADERS_TIMEOUT_MS:-300000}
      - STREAM_IDLE_TIMEOUT_MS=${STREAM_IDLE_TIMEOUT_MS:-600000}
      - RATE_LIMIT_EXECUTION_MAX_WAIT_MS=${RATE_LIMIT_EXECUTION_MAX_WAIT_MS:-300000}
      - RATE_LIMIT_MAX_WAIT_MS=${RATE_LIMIT_MAX_WAIT_MS:-300000}
      - OMNIROUTE_DIRECT_DISPATCHER_CONNECTIONS=${OMNIROUTE_DIRECT_DISPATCHER_CONNECTIONS:-32}
    ports:
      - "${APP_BIND_HOST:-0.0.0.0}:20128:20128"
    volumes:
      - omniroute-data:/app/data
    extra_hosts:
      - "host.docker.internal:host-gateway"
    restart: unless-stopped

volumes:
  omniroute-data: