Sync Gitbook Connection

File types to automatically sync when the data source connects. Only a subset of file types can be controlled. If not supported, then they will always be synced

Automatically sync attachments from files where supported. Currently applies to Helpdesk Tickets

Detect audio language before transcription for audio files

The transcription service to use for audio files. If no service is specified, 'deepgram' will be used.

Detect multiple speakers and label segments of speech by speaker for audio files.

Whether to split tabular rows into chunks. Currently only valid for CSV, TSV, and XLSX files.

If this flag is enabled, the file will be chunked and stored with Carbon, but no embeddings will be generated. This overrides the skip_embedding_generation flag.

If this flag is enabled, the file will be stored with Carbon, but no chunks or embeddings will be generated. This overrides the skip_embedding_generation and generate_chunks_only flags.

Setting this flag will create a new file record with Carbon but skip any and all processing. This means that we do not download the remote file content or generate any chunks or embeddings. We will store some metadata like name, external id, and external URL depending on the source you are syncing from. Note that this flag overrides both skip_embedding_generation and generate_chunks_only flags. The file will be moved to READY_TO_SYNC status.