compose.yaml

name: rememberstack

x-app: &app
  build:
    context: .
    dockerfile: Dockerfile
    args:
      # Stamped into the image so the benchmark can verify what actually ran.
      # Empty when the caller does not supply it; released images get the tag's
      # commit from CI.
      REMEMBERSTACK_BUILD_REVISION: ${REMEMBERSTACK_BUILD_REVISION:-}
  image: ghcr.io/writeitai/remember-stack:0.16.0
  environment:
    REMEMBERSTACK_DATABASE_URL: postgresql+psycopg://${REMEMBERSTACK_POSTGRES_USER}:${REMEMBERSTACK_POSTGRES_PASSWORD}@postgres:5432/${REMEMBERSTACK_POSTGRES_DB}
    REMEMBERSTACK_MINIO_ENDPOINT_URL: http://minio:9000
    REMEMBERSTACK_MINIO_ACCESS_KEY: ${REMEMBERSTACK_MINIO_ACCESS_KEY}
    REMEMBERSTACK_MINIO_SECRET_KEY: ${REMEMBERSTACK_MINIO_SECRET_KEY}
    REMEMBERSTACK_OPENROUTER_API_KEY: ${REMEMBERSTACK_OPENROUTER_API_KEY}
    REMEMBERSTACK_OPENROUTER_EMBEDDING_PROVIDER: ${REMEMBERSTACK_OPENROUTER_EMBEDDING_PROVIDER:-}
    REMEMBERSTACK_OPENROUTER_EMBEDDING_PROVIDER_ORDER: ${REMEMBERSTACK_OPENROUTER_EMBEDDING_PROVIDER_ORDER:-}
    REMEMBERSTACK_OPENROUTER_MAX_COMPLETION_TOKENS: ${REMEMBERSTACK_OPENROUTER_MAX_COMPLETION_TOKENS:-}
    REMEMBERSTACK_OPENROUTER_REASONING_EFFORT: ${REMEMBERSTACK_OPENROUTER_REASONING_EFFORT:-}
    REMEMBERSTACK_OPENROUTER_REASONING_EFFORT_MAP: ${REMEMBERSTACK_OPENROUTER_REASONING_EFFORT_MAP:-}
    REMEMBERSTACK_OPENROUTER_INVALID_COMPLETION_CAPTURE_DIR: ${REMEMBERSTACK_OPENROUTER_INVALID_COMPLETION_CAPTURE_DIR:-}
    REMEMBERSTACK_SELFHOST_DEPLOYMENT_ID: ${REMEMBERSTACK_SELFHOST_DEPLOYMENT_ID}
    REMEMBERSTACK_SELFHOST_DEPLOYMENT_SLUG: ${REMEMBERSTACK_SELFHOST_DEPLOYMENT_SLUG}
    REMEMBERSTACK_SELFHOST_DEPLOYMENT_NAME: ${REMEMBERSTACK_SELFHOST_DEPLOYMENT_NAME}
    REMEMBERSTACK_SELFHOST_API_PORT: "8000"
    # Believe X-Ingest-Principal-* on POST /ingest (D101). Leave false unless
    # this deployment is reachable ONLY by a control plane that already
    # authenticated the actor it names; on a public perimeter attribution
    # would be meaningless. Untrusted claims are ignored, never rejected.
    REMEMBERSTACK_SELFHOST_TRUSTED_PRINCIPAL_SOURCE: "${REMEMBERSTACK_SELFHOST_TRUSTED_PRINCIPAL_SOURCE:-false}"
    REMEMBERSTACK_SELFHOST_API_BEARER_BIND: ${REMEMBERSTACK_SELFHOST_API_BEARER_BIND:-}
    REMEMBERSTACK_SELFHOST_API_BEARER_TOKEN: ${REMEMBERSTACK_SELFHOST_API_BEARER_TOKEN:-}
    REMEMBERSTACK_SELFHOST_REQUIRE_API_AUTH: ${REMEMBERSTACK_SELFHOST_REQUIRE_API_AUTH:-false}
    # Browser origins allowed to call this deployment. Empty advertises no
    # CORS at all, which is the right default for a deployment no app points at.
    REMEMBERSTACK_SELFHOST_BROWSER_ORIGINS: ${REMEMBERSTACK_SELFHOST_BROWSER_ORIGINS:-}
    REMEMBERSTACK_SELFHOST_SPEND_LEASE_URL: ${REMEMBERSTACK_SELFHOST_SPEND_LEASE_URL:-}
    REMEMBERSTACK_SELFHOST_METER_INGEST_URL: ${REMEMBERSTACK_SELFHOST_METER_INGEST_URL:-}
    REMEMBERSTACK_SELFHOST_METER_INGEST_TOKEN: ${REMEMBERSTACK_SELFHOST_METER_INGEST_TOKEN:-}
    REMEMBERSTACK_SELFHOST_METER_IDENTITY_KEY: ${REMEMBERSTACK_SELFHOST_METER_IDENTITY_KEY:-}
    REMEMBERSTACK_SELFHOST_METER_ORG_ID: ${REMEMBERSTACK_SELFHOST_METER_ORG_ID:-}
    REMEMBERSTACK_SELFHOST_METER_PROJECT_ID: ${REMEMBERSTACK_SELFHOST_METER_PROJECT_ID:-}
    REMEMBERSTACK_SELFHOST_REQUIRE_METERING: ${REMEMBERSTACK_SELFHOST_REQUIRE_METERING:-false}
    REMEMBERSTACK_SELFHOST_CONVERSION_ROUTES: ${REMEMBERSTACK_SELFHOST_CONVERSION_ROUTES:-}
    REMEMBERSTACK_SELFHOST_INGEST_BODY_MAX_BYTES: ${REMEMBERSTACK_SELFHOST_INGEST_BODY_MAX_BYTES:-}
    REMEMBERSTACK_MISTRAL_OCR_API_KEY: ${REMEMBERSTACK_MISTRAL_OCR_API_KEY:-}
    REMEMBERSTACK_MISTRAL_OCR_BASE_URL: ${REMEMBERSTACK_MISTRAL_OCR_BASE_URL:-https://api.mistral.ai}
    REMEMBERSTACK_MISTRAL_OCR_MODEL: ${REMEMBERSTACK_MISTRAL_OCR_MODEL:-mistral-ocr-latest}
    REMEMBERSTACK_MISTRAL_OCR_TIMEOUT_S: ${REMEMBERSTACK_MISTRAL_OCR_TIMEOUT_S:-300}
    REMEMBERSTACK_MISTRAL_OCR_MAX_DOCUMENT_BYTES: ${REMEMBERSTACK_MISTRAL_OCR_MAX_DOCUMENT_BYTES:-50000000}
    REMEMBERSTACK_MISTRAL_OCR_INCLUDE_IMAGES: ${REMEMBERSTACK_MISTRAL_OCR_INCLUDE_IMAGES:-true}
    REMEMBERSTACK_MISTRAL_OCR_TABLE_FORMAT: ${REMEMBERSTACK_MISTRAL_OCR_TABLE_FORMAT:-markdown}
    REMEMBERSTACK_MISTRAL_OCR_EXTRACT_HEADERS_AND_FOOTERS: ${REMEMBERSTACK_MISTRAL_OCR_EXTRACT_HEADERS_AND_FOOTERS:-true}
    REMEMBERSTACK_MISTRAL_OCR_CONFIDENCE_GRANULARITY: ${REMEMBERSTACK_MISTRAL_OCR_CONFIDENCE_GRANULARITY:-word}
    REMEMBERSTACK_MISTRAL_OCR_KEEP_PROVIDER_RESPONSE: ${REMEMBERSTACK_MISTRAL_OCR_KEEP_PROVIDER_RESPONSE:-true}
    REMEMBERSTACK_MISTRAL_OCR_PRICE_USD_PER_1000_PAGES: ${REMEMBERSTACK_MISTRAL_OCR_PRICE_USD_PER_1000_PAGES:-1}
    REMEMBERSTACK_SELFHOST_GRAPH_POOL_SIZE: ${REMEMBERSTACK_SELFHOST_GRAPH_POOL_SIZE:-4}
    REMEMBERSTACK_SELFHOST_GRAPH_POOL_TIMEOUT_S: ${REMEMBERSTACK_SELFHOST_GRAPH_POOL_TIMEOUT_S:-1}
    REMEMBERSTACK_SELFHOST_GRAPH_MAX_CONCURRENCY: ${REMEMBERSTACK_SELFHOST_GRAPH_MAX_CONCURRENCY:-2}
    REMEMBERSTACK_SELFHOST_GRAPH_WORK_MEM_KIB: ${REMEMBERSTACK_SELFHOST_GRAPH_WORK_MEM_KIB:-16384}
    REMEMBERSTACK_SELFHOST_RETRIEVAL_POOL_SIZE: ${REMEMBERSTACK_SELFHOST_RETRIEVAL_POOL_SIZE:-4}
    REMEMBERSTACK_SELFHOST_RETRIEVAL_POOL_TIMEOUT_S: ${REMEMBERSTACK_SELFHOST_RETRIEVAL_POOL_TIMEOUT_S:-1}
    REMEMBERSTACK_SELFHOST_RETRIEVAL_MAX_CONCURRENCY: ${REMEMBERSTACK_SELFHOST_RETRIEVAL_MAX_CONCURRENCY:-4}
    REMEMBERSTACK_SENTRY_DSN: ${REMEMBERSTACK_SENTRY_DSN:-}
    REMEMBERSTACK_SENTRY_ENVIRONMENT: ${REMEMBERSTACK_SENTRY_ENVIRONMENT:-}
    REMEMBERSTACK_SENTRY_SAMPLE_RATE: ${REMEMBERSTACK_SENTRY_SAMPLE_RATE:-}
    LANGFUSE_PUBLIC_KEY: ${LANGFUSE_PUBLIC_KEY:-}
    LANGFUSE_SECRET_KEY: ${LANGFUSE_SECRET_KEY:-}
    LANGFUSE_HOST: ${LANGFUSE_HOST:-}
    # D79: STRUCTURER now names only the anchor-proposal fallback seat.
    REMEMBERSTACK_STRUCTURER_MODEL: ${REMEMBERSTACK_STRUCTURER_MODEL:-openai/gpt-5.6-luna}
    REMEMBERSTACK_SKELETON_CHECK_MODEL: ${REMEMBERSTACK_SKELETON_CHECK_MODEL:-z-ai/glm-4.7-flash}
    REMEMBERSTACK_ROLE_MODEL: ${REMEMBERSTACK_ROLE_MODEL:-z-ai/glm-4.7-flash}
    REMEMBERSTACK_SUMMARY_MODEL: ${REMEMBERSTACK_SUMMARY_MODEL:-z-ai/glm-4.7-flash}
    REMEMBERSTACK_E1_EMBEDDING_MODEL: ${REMEMBERSTACK_E1_EMBEDDING_MODEL:-qwen/qwen3-embedding-8b}
    REMEMBERSTACK_E1_PREFIX_MODEL: ${REMEMBERSTACK_E1_PREFIX_MODEL:-openai/gpt-5.6-luna}
    REMEMBERSTACK_E2_EXTRACT_MODEL: ${REMEMBERSTACK_E2_EXTRACT_MODEL:-openai/gpt-5.6-luna}
    REMEMBERSTACK_E3_NORMALIZE_MODEL: ${REMEMBERSTACK_E3_NORMALIZE_MODEL:-openai/gpt-5.6-luna}
    REMEMBERSTACK_OBS_SMALL_MODEL: ${REMEMBERSTACK_OBS_SMALL_MODEL:-openai/gpt-5.6-luna}
    REMEMBERSTACK_OBS_FRONTIER_MODEL: ${REMEMBERSTACK_OBS_FRONTIER_MODEL:-openai/gpt-5.6-sol}
    REMEMBERSTACK_OBS_EMBEDDING_MODEL: ${REMEMBERSTACK_OBS_EMBEDDING_MODEL:-qwen/qwen3-embedding-8b}
    REMEMBERSTACK_ADJUDICATOR_SMALL_MODEL: ${REMEMBERSTACK_ADJUDICATOR_SMALL_MODEL:-openai/gpt-5.6-luna}
    REMEMBERSTACK_ADJUDICATOR_FRONTIER_MODEL: ${REMEMBERSTACK_ADJUDICATOR_FRONTIER_MODEL:-openai/gpt-5.6-sol}
    REMEMBERSTACK_P1_EMBEDDING_MODEL: ${REMEMBERSTACK_P1_EMBEDDING_MODEL:-qwen/qwen3-embedding-8b}
    REMEMBERSTACK_P1_LABEL_MODEL: ${REMEMBERSTACK_P1_LABEL_MODEL:-openai/gpt-5.6-luna}
  volumes:
    - app-state:/var/lib/rememberstack
    - forget-manifests:/var/lib/rememberstack/forget-manifests

services:
  postgres:
    build:
      context: .
      dockerfile: Dockerfile.postgres
    image: rememberstack-postgres:19beta3
    command:
      - postgres
      - -c
      - shared_preload_libraries=pg_textsearch,pg_partman_bgw
      - -c
      - pg_partman_bgw.dbname=${REMEMBERSTACK_POSTGRES_DB}
      - -c
      - pg_partman_bgw.role=${REMEMBERSTACK_POSTGRES_USER}
    environment:
      POSTGRES_USER: ${REMEMBERSTACK_POSTGRES_USER}
      POSTGRES_PASSWORD: ${REMEMBERSTACK_POSTGRES_PASSWORD}
      POSTGRES_DB: ${REMEMBERSTACK_POSTGRES_DB}
    healthcheck:
      test:
        - CMD-SHELL
        - >-
          psql --username ${REMEMBERSTACK_POSTGRES_USER}
          --dbname ${REMEMBERSTACK_POSTGRES_DB}
          --tuples-only --command 'SELECT 1' >/dev/null
      interval: 2s
      timeout: 5s
      retries: 30
    volumes:
      # Official PostgreSQL 18+ images keep major-version directories below
      # this mount so pg_upgrade can link across a version boundary.
      - postgres-data:/var/lib/postgresql

  minio:
    image: quay.io/minio/minio:RELEASE.2025-09-07T16-13-09Z@sha256:14cea493d9a34af32f524e538b8346cf79f3321eff8e708c1e2960462bd8936e
    command: server /data --console-address :9001
    environment:
      MINIO_ROOT_USER: ${REMEMBERSTACK_MINIO_ACCESS_KEY}
      MINIO_ROOT_PASSWORD: ${REMEMBERSTACK_MINIO_SECRET_KEY}
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
      interval: 2s
      timeout: 5s
      retries: 30
    ports:
      - "9001:9001"
    volumes:
      - minio-data:/data

  setup:
    <<: *app
    command: ["setup"]
    depends_on:
      postgres:
        condition: service_healthy
      minio:
        condition: service_healthy

  api:
    <<: *app
    command: ["api"]
    depends_on:
      setup:
        condition: service_completed_successfully
    healthcheck:
      test:
        - CMD
        - python
        - -c
        - "import urllib.request; urllib.request.urlopen('http://localhost:8000/healthz', timeout=2).read()"
      interval: 10s
      timeout: 5s
      retries: 30
    ports:
      - "${REMEMBERSTACK_SELFHOST_API_PORT}:8000"

  worker-convert:
    <<: *app
    command: ["worker", "--stage", "convert"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-structure:
    <<: *app
    command: ["worker", "--stage", "structure"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-chunk:
    <<: *app
    command: ["worker", "--stage", "chunk"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-embed-chunk:
    <<: *app
    command: ["worker", "--stage", "embed_chunk"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-extract-claims:
    <<: *app
    command: ["worker", "--stage", "extract_claims"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-normalize-relations:
    <<: *app
    command: ["worker", "--stage", "normalize_relations"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-adjudicate-observations:
    <<: *app
    command: ["worker", "--stage", "adjudicate_observations"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-adjudicate-supersession:
    <<: *app
    command: ["worker", "--stage", "adjudicate_supersession"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-embed-claim:
    <<: *app
    command: ["worker", "--stage", "embed_claim"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-reconcile:
    <<: *app
    command: ["worker", "--stage", "reconcile"]
    depends_on:
      setup:
        condition: service_completed_successfully

  worker-label-relation:
    <<: *app
    command: ["worker", "--stage", "label_relation"]
    depends_on:
      setup:
        condition: service_completed_successfully

  meter-receipts:
    <<: *app
    command: ["meter-receipts"]
    profiles: ["managed"]
    depends_on:
      setup:
        condition: service_completed_successfully

  projections:
    <<: *app
    command: ["project", "--plane", "p3"]
    profiles: ["operations"]
    depends_on:
      setup:
        condition: service_completed_successfully

volumes:
  app-state:
  forget-manifests:
  minio-data:
  postgres-data: