P4 Search configuration reference

The product version that is shown in the P4 Server log. Default value: the installed version of P4 Search.

Set the file name for the P4 Search log.

The host IP address for the P4 Search service.

The port number for the P4 Search service.

The protocol used for the P4 Search service. Public to machines on the same network by default. Options are http or https.

The location of the Java Keystore. Only applies if protocol is set to https. See Secure P4 Search communications.

The location of the Java Keystore. Only applies if protocol is set to https. See Secure P4 Search communications.

Location of the P4 Search local trust store.

Password for the P4 Search local trust store. This password is randomly generated.

Specify origins for the browser to allow resource access. This type of permission is also known as Cross Origin Resource Sharing (CORS).

Include your domain along with the protocol and port number. Comma-separated values are supported. For example, https://*.mydomain.com:1601,https://*.mydomain.com:1602

If the CORS value is empty (the default), the external URL value is used as an allowed origin along with the protocol and port number. To secure your site, you must use https as the protocol.

The retry interval used by P4 Search when locating Elasticsearch and P4 Server Services (in milliseconds).

An X-Auth-Token to validate configuration requests.

The URL of the P4 Search help documentation. Changing this URL is useful if, for example, you are running an isolated system and need to host it on your own server.

The URL of your P4 Search instance as seen by other applications. For example, this external URL can be used by P4 Server or P4 DAM to connect to a specific instance of P4 Search.

The number of live log lines displayed in the logging window of the P4 Search web UI.

The number of requests allowed per minute for all API endpoints.

The number of login requests allowed per minute for all API endpoints.

The duration after which the JWT Access token expires.

Ensure that the value of the JWT access token is always smaller than the value of the JWT refresh token.

The duration (in seconds) after which the JWT Refresh token expires.

Ensure that the value of the JWT refresh token is always greater than the value of the JWT access token.

A nonce token is a unique, random number or string that is used only once and then discarded. It acts as a security measure to prevent replay attacks in communication protocols, particularly in API requests.

The nonce token is included in the request header along with the authentication token, and it must be present in all API requests for security purposes. This practice ensures that each request is unique, even when the same authentication details are used.

To enable a nonce token, which is a unique string, set com.perforce.p4search.service.nonce to true.

The P4 Server RPC host and port. For example, perforce.com:1666, or ssl.perforce.com:1666 if protocol is set to https.

The p4trust trusted fingerprint.

A comma-separated list of trusted IP addresses for proxy services. For example, the IP address for your P4 Code Review instance.

If you are using P4 Search with another service, such as P4 Code Review, and are using P4 Server Security level 5 and higher, provide a list of trusted IP addresses for the services. For example:

com.perforce.p4search.core.allowed=swarm.perforce.com,proxy.perforce.com

Can be a Standard or Service user type with a minimum of admin access. Used to authorize the proxy connection at Security level 5 and above. To learn more about creating the Service user and printing out the ticket, see Create users and tickets for external services.

The long-lived ticket to authorize the Perforce Service User. To print out a long-lived ticket for the Perforce service user that is valid for all hosts, run the following command : p4 -p <p4port> -u <p4user> login -a -p.

Optional:Must be a Standard user type with super access. Used to read content from the P4 Server, to set file attributes, and to install extensions. For instructions on creating the Index user and printing out the ticket, see Create users and tickets for external services.

The long-lived ticket to authorize the index user. To print out a long-lived ticket for the index user that is valid for all hosts, run the following command : p4 -p <p4port> -u <p4user> login -a -p.

A flag to automatically update extensions during start-up.

Set to true (the default) to generate the blur and thumbnail/preview images when indexing is run.

• Blur images are stored on the P4 Server as an attribute of the original image file and indexed in Elasticsearch. They consist of a 4-pixel image stored as a Hash String that loads extremely quickly. P4 DAMdisplays the blur images for search queries while the thumbnail/preview images are loaded.

• Thumbnails/previews are stored on the P4 Server as an attribute of the original image file. They are up to 240 x 240 pixels depending on the original image's aspect ratio. P4V and P4 DAM use thumbnails for image previews.

Set to Enable to enable Tesseract OCR.

A comma-separated list of host addresses for Elasticsearch. For example, http://localhost:9200,http://localhost:9201

To configure Elasticsearch hosts, P4 Search must have access to Elasticsearch to create and update the alias index for the P4 Server data. P4 Search creates a partitioned index.

Elasticsearch user for P4 Search. Use to configure Elasticsearch authentication.

Elasticsearch password for P4 Search. Use to configure Elasticsearch authentication.

To filter user search results based on their P4 Server permissions (output of the p4 protect command), set to true.

This requires the installation of the P4 Serverp4search-filter plugin for Elasticsearch. See the Elasticsearch plugin documentation.

Use of the P4 Serverp4search-filter plugin introduces a small overhead in the query time. This is mainly determined by the number of protection entries that are applied to the user running the query. In some situations, you can disable the filter if P4 Search contains only indexed data that is visible to all users.

Maximum accurately counted total hit by Elasticsearch. To tune performance, set a limit to the number of results processed by a single query.

Timeout (in milliseconds) when parsing a file through Tika.

The model used for auto-tagging images. An empty value (the default) disables image auto-tagging.

Supported models include:

• ClipModel (OpenAI)

• OpenAiDetectModel (OpenAI)

• AzureTagsModel (Azure)

• RekognitionLabelsModel (AWS)

• GoogleLabelModel (Google)

• DeepDetectModel (DeepDetect)

The host IP address for the auto-tagging service. Configurable shared with the Configure image auto-OCR configurable.

The language you to generate the image tags in. For example, en (English). For supported languages, see the documentation for your auto-tagging service. Configurable shared with the Configure image auto-OCR configurable.

The API key for the auto-tagging service you are using. Configurable shared with the Configure image auto-OCR configurable.

The automatic image detection threshold as a floating point percentage. For example, 0.1=10%.

The automatic best image detection results limit.

The classification service used for image files.

The following OpenAI services are supported:

• gpt-5-nano

• gpt-5-mini

• gpt-4.1-mini

• o4-mini

The following P4 Detect models are supported:

• clip_model

The following DeepDetect services are supported:

• classification_21k

• imageserv

To learn more about using a DeepDetect model, see Setup of an image classifier from the DeepDetect server documentation.

The model used for automated speech-to-text for audio and video files. An empty value (the default) disables automated speech-to-text.

Valid models include:

• AzureSpeechModel

• AmazonTranscribeModel

• GoogleSpeechModel

• WhisperSpeechModel

The host IP address for the auto-speech service.

The language of the audio you want to be transcribed.

This configurable is optional for WhisperSpeechModel.

The API key for the auto-speech recognition service you are using.

This configurable is optional for WhisperSpeechModel.

Set the automatic speech recognition threshold for your auto-speech service as a floating point percentage. The automatic speech recognition threshold is the accuracy returned by your auto-speech service. For example, 0.6=60%.

In this example, when auto-speech threshold is set to 0.6, any transcription with 60% accuracy is accepted by the auto-speech service and anything below that is discarded.

Set the audio transcribe limit for your auto-speech service in seconds. The default audio transcribe limit is set to 60 seconds and anything longer than that is discarded.

The name of the cloud storage bucket or container where the large audio files are stored. This is only available for AmazonTranscribeModel.

The render service implementation. For example, AWSRenderService or LocalRenderService.

Deprecated.

The rendering service API security key for use with AWSRenderService. The LocalRenderService uses the X-Auth-Token configured in Index auth-token.

Checks if the external URL of your P4 Search instance is reachable.

Indexes your assets.

Indexes your changes.