Skip to main content
Applies to BloodHound Enterprise and CE In BloodHound Community Edition, you can run OpenHound as a standalone CLI application to execute built-in collectors or community-built collectors. You can configure OpenHound with environment variables, configuration files, or both. See the configuration page to set up OpenHound.

​
Installing the OpenHound CLI

OpenHound is typically installed and executed inside a dedicated Python virtual environment. We recommend using a Python package and project manager such as uv to manage the virtual environment and dependencies. When using uv, you can install OpenHound by running the following command in your terminal: 
uv tool install openhound
The core installation of OpenHound does not include any collectors. However, the following collectors are available as additional packages that can be installed along with the core OpenHound package:
  • Okta
  • Jamf
  • Github
You can specify which collector(s) to include by adding the collector name as an extra dependency with the install command. You can specify one or more collectors by separating them with a comma. For example, to install OpenHound with the Okta, Jamf, and Github collectors, run the following command: 
uv tool install "openhound[okta,jamf,github]"
Use the openhound[all] alias to install all the built-in collectors.

​
CLI Commands

The openhound command supports several subcommands for running collectors and converting data to the OpenGraph format. The following sections provide an overview of the commonly used subcommands. For a full list of available commands and options, refer to the built-in help documentation by running openhound --help or openhound <subcommand> --help.

​
Collect

The collect subcommand is used to start the collection process. You can start collecting data by specifying the collector name and output directory. For example, to run the Okta collector and save the output in the output directory: 
openhound collect okta ./output
OpenHound will then run the Okta collector and save the collected resources in the ./output/okta directory. The collected resources will be saved in a JSONL compressed format and can be processed and converted to OpenGraph using the preproc and convert subcommands.

​
Preprocess

The preproc subcommand preprocesses collected resources before they are converted into OpenGraph data. During this phase, OpenHound imports raw data into a DuckDB database and creates materialized tables to support efficient lookups for transformations. This phase is optional and is only required for collectors that depend on preprocessed data. The following command preprocesses Okta resources from the ./output/okta directory and generates a DuckDB lookup database at ./lookup.duckdb: 
openhound preproc okta ./output/okta ./lookup.duckdb
The preprocessed data can then be used as a lookup database in the convert phase to generate the OpenGraph output (if required).

​
Convert

The convert subcommand transforms collected resources into the OpenGraph format. This phase converts the raw data to nodes and edges and saves them as OpenGraph-compatible JSON files that you can upload to BloodHound. The following command converts the collected Okta resources from the ./output/okta directory and saves the OpenGraph files in the ./graph/okta directory: 
openhound convert okta ./output/okta ./graph/okta
OpenHound will automatically split each graph into multiple files based on source node and chunk size to optimize file uploads for BloodHound.