S3 Connector
Connector Configuration
The S3 connector allows you to synchronize files and documents stored in an S3-compatible bucket (AWS S3, OVH, Scaleway, or any other compatible provider) to Wikit Semantics.
| Field Name | Format / Type | Required | Comment |
|---|---|---|---|
| S3 Endpoint URL | URL | ✅ | URL of your S3 provider's access point (e.g., https://s3.gra.io.cloud.ovh.net) |
| Bucket Name | Free text | ✅ | Name of the S3 bucket to synchronize |
| Access Key ID | Free text | ✅ | Access Key ID for authentication |
| Secret Access Key | Password / Token | ✅ | Secret Access Key associated with the Access Key ID |
| Region | Free text | ➖ | Region ID required by some providers (e.g., gra for OVH, fr-par for Scaleway) |
| Prefix | Free text | ➖ | Base path to limit synchronization to a subfolder of the bucket (e.g., documents/public/) |
| Paths to include | List of texts | ➖ | Limits synchronization to the listed folders. Expected format: Library-name/Folder-name |
| Paths to exclude | List of texts | ➖ | Excludes the listed folders from synchronization. Expected format: Library-name/Folder-name |
| Force path-style mode | Yes / No | ➖ | Enabled by default. Disable only if your provider requires virtual-hosted mode |
| Disable source URL | Yes / No | ➖ | When enabled, no source URL will be associated with synchronized documents |
| Source URL metadata field | Free text | ➖ | Name of the S3 metadata field containing the source URL to associate with the document |
| Modification date metadata field | Free text | ➖ | Name of the S3 metadata field containing the modification date (ISO format). If absent or invalid, the native S3 object date is used |
💡 Fields marked ✅ are required for the connector to work.
How the Connector Works
The S3 connector operates in smart synchronization mode:
- Change detection: it inserts, updates, or deletes documents by comparing the current state of the bucket with the state already indexed in Wikit Semantics.
- Modification detection: the modification date is retrieved from the configured metadata field (
Modification date metadata field), in ISO format. If this field is absent or invalid, the native S3 object modification date is used instead. - Path filters:
Prefixlimits synchronization to a subfolder of the bucket.Paths to includeandPaths to excludeallow fine-tuning the folders to synchronize. - Source URL: by default, a source URL is associated with each synchronized document. You can disable this association (
Disable source URL) or customize it by specifying the S3 metadata field containing the URL to use (Source URL metadata field).
FAQ
How do I find the S3 endpoint URL?
The endpoint URL depends on your provider. Some examples:
- OVH:
https://s3.<region>.io.cloud.ovh.net - Scaleway:
https://s3.<region>.scw.cloud - AWS:
https://s3.<region>.amazonaws.com
Refer to your provider's documentation to find the exact endpoint associated with your bucket.
How do I generate an access key (Access Key / Secret Key)?
Each provider has its own access key management space (IAM or equivalent). You must create or retrieve:
- an Access Key ID
- a Secret Access Key
These credentials must have at least read rights on the bucket to synchronize.
What is the difference between path-style and virtual-hosted mode?
- Path-style (default): the bucket access URL takes the form
https://endpoint/bucket/object. Compatible with most S3 providers. - Virtual-hosted: the URL takes the form
https://bucket.endpoint/object. Required by some providers or specific configurations.
If your provider supports and requires virtual-hosted mode, disable the Force path-style mode option.
How do I limit synchronization to a subfolder of the bucket?
Use the Prefix field to specify the base path (e.g., documents/public/). Only objects whose name starts with this prefix will be synchronized. For finer targeting, you can combine Prefix with the Paths to include and Paths to exclude fields.
Are documents deleted from the S3 bucket also deleted in Wikit Semantics?
Yes. During each synchronization cycle, the connector compares the state of the bucket with the already indexed documents. Any document absent from the bucket is automatically removed from Wikit Semantics.