config.yaml

# This file is automatically generated from build.yaml
# by build/gen_config.py - do not edit manually.
#
# Stripped provider types (dependency only - available via custom config.yaml):
#   - inline::docling
#   - inline::faiss
#   - inline::markitdown
#   - inline::milvus
#   - inline::pypdf
#   - inline::sentence-transformers
#   - remote::docling-serve
version: 2
image_name: rh
apis:
  - responses
  - messages
  - batches
  - inference
  - tool_runtime
  - vector_io
  - files
  - file_processors
providers:
  inference:
    - provider_id: ${env.VLLM_URL:+vllm-inference}
      provider_type: remote::vllm
      config:
        base_url: ${env.VLLM_URL:=}
        max_tokens: ${env.VLLM_MAX_TOKENS:=4096}
        api_token: ${env.VLLM_API_TOKEN:=fake}
        tls_verify: ${env.VLLM_TLS_VERIFY:=true}
        refresh_models: ${env.VLLM_REFRESH_MODELS:=false}
    - provider_id: ${env.VLLM_EMBEDDING_URL:+vllm-embedding}
      provider_type: remote::vllm
      config:
        base_url: ${env.VLLM_EMBEDDING_URL:=}
        max_tokens: ${env.VLLM_EMBEDDING_MAX_TOKENS:=4096}
        api_token: ${env.VLLM_EMBEDDING_API_TOKEN:=fake}
        tls_verify: ${env.VLLM_EMBEDDING_TLS_VERIFY:=true}
        refresh_models: ${env.VLLM_EMBEDDING_REFRESH_MODELS:=false}
    - provider_id: ${env.ENABLE_BEDROCK:+bedrock}
      provider_type: remote::bedrock
      config:
        aws_bedrock_bearer_token: ${env.AWS_BEDROCK_BEARER_TOKEN:=}
        region_name: ${env.AWS_DEFAULT_REGION:=us-east-2}
        aws_role_arn: ${env.AWS_ROLE_ARN:=}
        aws_web_identity_token_file: ${env.AWS_WEB_IDENTITY_TOKEN_FILE:=}
    - provider_id: ${env.WATSONX_API_KEY:+watsonx}
      provider_type: remote::watsonx
      config:
        base_url: ${env.WATSONX_BASE_URL:=https://us-south.ml.cloud.ibm.com}
        api_key: ${env.WATSONX_API_KEY:=}
        project_id: ${env.WATSONX_PROJECT_ID:=}
    - provider_id: ${env.AZURE_API_KEY:+azure}
      provider_type: remote::azure
      config:
        api_key: ${env.AZURE_API_KEY:=}
        api_base: ${env.AZURE_API_BASE:=}
        api_version: ${env.AZURE_API_VERSION:=}
        api_type: ${env.AZURE_API_TYPE:=}
    - provider_id: ${env.VERTEX_AI_PROJECT:+vertexai}
      provider_type: remote::vertexai
      config:
        project: ${env.VERTEX_AI_PROJECT:=}
        location: ${env.VERTEX_AI_LOCATION:=global}
    - provider_id: ${env.OPENAI_API_KEY:+openai}
      provider_type: remote::openai
      config:
        api_key: ${env.OPENAI_API_KEY:=}
        base_url: ${env.OPENAI_BASE_URL:=https://api.openai.com/v1}
    - provider_id: ${env.ENABLE_GEMINI:+gemini}
      provider_type: remote::gemini
      config:
        api_key: ${env.GEMINI_API_KEY:=}
        access_token: ${env.GEMINI_ACCESS_TOKEN:=}
        project: ${env.GEMINI_AI_PROJECT:=}
    - provider_id: ${env.ANTHROPIC_API_KEY:+anthropic}
      provider_type: remote::anthropic
      config:
        api_key: ${env.ANTHROPIC_API_KEY:=}
  vector_io:
    - provider_id: ${env.MILVUS_ENDPOINT:+milvus-remote}
      provider_type: remote::milvus
      config:
        uri: ${env.MILVUS_ENDPOINT:=}
        token: ${env.MILVUS_TOKEN:=}
        secure: ${env.MILVUS_SECURE:=}
        consistency_level: ${env.MILVUS_CONSISTENCY_LEVEL:=}
        ca_pem_path: ${env.MILVUS_CA_PEM_PATH:=}
        client_pem_path: ${env.MILVUS_CLIENT_PEM_PATH:=}
        client_key_path: ${env.MILVUS_CLIENT_KEY_PATH:=}
        persistence:
          backend: kv_default
          namespace: vector_io::milvus_remote
    - provider_id: ${env.ENABLE_PGVECTOR:+pgvector}
      provider_type: remote::pgvector
      config:
        host: ${env.PGVECTOR_HOST:=localhost}
        port: ${env.PGVECTOR_PORT:=5432}
        db: ${env.PGVECTOR_DB:=}
        user: ${env.PGVECTOR_USER:=}
        password: ${env.PGVECTOR_PASSWORD:=}
        persistence:
          backend: kv_default
          namespace: vector_io::pgvector
    - provider_id: ${env.ENABLE_QDRANT:+qdrant-remote}
      provider_type: remote::qdrant
      config:
        location: ${env.QDRANT_LOCATION:=}
        url: ${env.QDRANT_URL:=}
        port: ${env.QDRANT_PORT:=6333}
        grpc_port: ${env.QDRANT_GRPC_PORT:=6334}
        prefer_grpc: ${env.QDRANT_PREFER_GRPC:=false}
        https: ${env.QDRANT_HTTPS:=}
        api_key: ${env.QDRANT_API_KEY:=}
        prefix: ${env.QDRANT_PREFIX:=}
        timeout: ${env.QDRANT_TIMEOUT:=}
        host: ${env.QDRANT_HOST:=}
        persistence:
          backend: kv_default
          namespace: vector_io::qdrant_remote
  responses:
    - provider_id: meta-reference
      provider_type: inline::builtin
      config:
        persistence:
          agent_state:
            backend: kv_default
            namespace: agents::meta_reference
          responses:
            backend: sql_default
            table_name: agents_responses
            max_write_queue_size: 10000
            num_writers: 4
  messages:
  # Implements the Anthropic Messages API (/v1/messages). Native passthrough for
  # providers that support it (e.g. vLLM); translates Anthropic <-> OpenAI Chat
  # Completions for all other providers. Depends on the inference API.
    - provider_id: builtin
      provider_type: inline::builtin
      config:
        kvstore:
          backend: kv_default
          namespace: message_batches
  tool_runtime:
  # NOTE: inline::file-search requires a 'files' provider at runtime
  # (currently inline::localfs). Do not add inline::localfs to
  # STRIPPED_PROVIDER_TYPES in gen_config.py without also removing
  # inline::file-search here.
    - provider_id: brave-search
      provider_type: remote::brave-search
      config:
        api_key: ${env.BRAVE_SEARCH_API_KEY:=}
        max_results: 3
    - provider_id: tavily-search
      provider_type: remote::tavily-search
      config:
        api_key: ${env.TAVILY_SEARCH_API_KEY:=}
        max_results: 3
    - provider_id: file-search
      provider_type: inline::file-search
      config: {}
    - provider_id: model-context-protocol
      provider_type: remote::model-context-protocol
      config: {}
  files:
  # Required at runtime by inline::file-search (see tool_runtime above).
    - provider_id: meta-reference-files
      provider_type: inline::localfs
      config:
        storage_dir: /opt/app-root/src/.ogx/distributions/rh/files
        metadata_store:
          backend: sql_default
          table_name: files_metadata
    - provider_id: ${env.ENABLE_S3:+s3}
      provider_type: remote::s3
      config:
        bucket_name: ${env.S3_BUCKET_NAME:=}
        region: ${env.AWS_DEFAULT_REGION:=us-east-1}
        aws_access_key_id: ${env.AWS_ACCESS_KEY_ID:=}
        aws_secret_access_key: ${env.AWS_SECRET_ACCESS_KEY:=}
        endpoint_url: ${env.S3_ENDPOINT_URL:=}
        auto_create_bucket: ${env.S3_AUTO_CREATE_BUCKET:=false}
        metadata_store:
          backend: sql_default
          table_name: files_metadata
  file_processors:
    - provider_id: auto
      provider_type: inline::auto
      config: {}
  batches:
    - provider_id: reference
      provider_type: inline::reference
      config:
        sqlstore:
          table_name: batches
          backend: sql_default
storage:
  backends:
    kv_default:
      type: kv_postgres
      host: ${env.POSTGRES_HOST:=localhost}
      port: ${env.POSTGRES_PORT:=5432}
      db: ${env.POSTGRES_DB:=}
      user: ${env.POSTGRES_USER:=}
      password: ${env.POSTGRES_PASSWORD:=}
      table_name: ${env.POSTGRES_TABLE_NAME:=ogx_kvstore}
    sql_default:
      type: sql_postgres
      host: ${env.POSTGRES_HOST:=localhost}
      port: ${env.POSTGRES_PORT:=5432}
      db: ${env.POSTGRES_DB:=}
      user: ${env.POSTGRES_USER:=}
      password: ${env.POSTGRES_PASSWORD:=}
  stores:
    metadata:
      backend: kv_default
      namespace: registry
    inference:
      table_name: inference_store
      backend: sql_default
      max_write_queue_size: 10000
      num_writers: 4
    conversations:
      table_name: openai_conversations
      backend: sql_default
    prompts:
      namespace: prompts
      backend: kv_default
    connectors:
      table_name: connectors
      backend: sql_default
    vector_stores:
      backend: sql_default
      table_name: vector_store_metadata
registered_resources:
  models:
    - metadata:
        embedding_dimension: ${env.EMBEDDING_DIMENSION:=768}
      model_id: ${env.EMBEDDING_MODEL:=}
      provider_id: ${env.EMBEDDING_PROVIDER:=vllm-embedding}
      provider_model_id: ${env.EMBEDDING_PROVIDER_MODEL_ID:=ibm-granite/granite-embedding-125m-english}
      model_type: embedding
  vector_dbs: []
  tool_groups:
    - toolgroup_id: builtin::websearch
      provider_id: tavily-search
    - toolgroup_id: builtin::rag
      provider_id: file-search
vector_stores:
  annotation_prompt_params:
    enable_annotations: true
    annotation_instruction_template: "Cite sources immediately at the end of sentences using <|file-id|> format."
telemetry:
  enabled: true
server:
  port: 8321
  registry_refresh_interval_seconds: ${env.REGISTRY_REFRESH_INTERVAL_SECONDS:=300}
  auth:
    provider_config:
      type: ${env.AUTH_ISSUER:+oauth2_token}
      audience: ${env.AUTH_AUDIENCE:=ogx}
      issuer: ${env.AUTH_ISSUER:=}
      jwks:
        uri: ${env.AUTH_JWKS_URI:=}
        key_recheck_period: ${env.AUTH_JWKS_RECHECK_PERIOD:=3600}
      verify_tls: ${env.AUTH_VERIFY_TLS:=true}
      tls_cafile: ${env.AUTH_TLS_CAFILE:=}
    access_policy:
      - permit:
          actions: [read]
        when: resource is unowned
        description: "All users can read system resources"
      - permit:
          actions: [create]
        description: "Authenticated users can create resources"
      - permit:
          actions: [read, update, delete]
        when: user is owner
        description: "Owners can manage their own resources"