Environment Variables for NeMo Retriever Library
The following are the environment variables that you can use to configure NeMo Retriever Library. You can specify these in your .env file or directly in your environment.
Note
NVIDIA Ingest (nv-ingest) has been renamed NeMo Retriever Library.
General Environment Variables
| Name | Example | Description |
|---|---|---|
DOWNLOAD_LLAMA_TOKENIZER |
- | The Llama tokenizer is now pre-downloaded at build time. For details, refer to Token-Based Splitting. |
HF_ACCESS_TOKEN |
- | A token to access HuggingFace models. For details, refer to Token-Based Splitting. |
INGEST_LOG_LEVEL |
- DEBUG - INFO - WARNING - ERROR - CRITICAL |
The log level for the ingest service, which controls the verbosity of the logging output. |
MESSAGE_CLIENT_HOST |
- redis - localhost - 192.168.1.10 |
Specifies the hostname or IP address of the message broker used for communication between services. |
MESSAGE_CLIENT_PORT |
- 7670 - 6379 |
Specifies the port number on which the message broker is listening. |
MINIO_BUCKET |
nv-ingest |
Name of MinIO bucket, used to store image, table, and chart extractions. |
NGC_API_KEY |
nvapi-************* |
An authorized NGC API key, used to interact with hosted NIMs. To create an NGC key, go to https://org.ngc.nvidia.com/setup/api-keys. |
NIM_NGC_API_KEY |
— | The key that NIM microservices inside docker containers use to access NGC resources. This is necessary only in some cases when it is different from NGC_API_KEY. If this is not specified, NGC_API_KEY is used to access NGC resources. |
OTEL_EXPORTER_OTLP_ENDPOINT |
http://otel-collector:4317 |
The endpoint for the OpenTelemetry exporter, used for sending telemetry data. |
REDIS_INGEST_TASK_QUEUE |
ingest_task_queue |
The name of the task queue in Redis where tasks are stored and processed. |
REDIS_POOL_SIZE |
- 50 (default) - 100 - 200 |
Maximum Redis connection pool size. Increase for high-concurrency workloads processing many documents in parallel. Default of 50 works well for most deployments. |
IMAGE_STORAGE_URI |
s3://nv-ingest/artifacts/store/images |
Default fsspec-compatible URI for the store task. Supports s3://, file://, gs://, etc. See Store Extracted Images. |
IMAGE_STORAGE_PUBLIC_BASE_URL |
https://assets.example.com/images |
Optional HTTP(S) base URL for serving stored images. |
Vector Database (Retrieval) Environment Variables
These variables apply when using the test harness or when configuring the vector database backend.
| Name | Example | Description |
|---|---|---|
VDB_BACKEND |
lancedb (default) milvus |
Vector database backend. Use lancedb for embedded, in-process storage (default), or milvus for client-server. |
HYBRID |
true false (default) |
LanceDB only: enable hybrid retrieval (BM25 FTS + vector, RRF). |
Library Mode Environment Variables
These environment variables apply specifically when running the library in library mode.
| Name | Example | Description |
|---|---|---|
NVIDIA_API_KEY |
nvapi-************* |
API key for NVIDIA-hosted NIM services. |