Docling

Process documents using Docling library for parsing and conversion.

Metadata

Property	Value
Scheme	`docling`
Support Level	Stable
Labels	transformation,ai
Version	4.20.0 (CEQ 3.36.0)

Maven Dependency

<dependency>
    <groupId>org.apache.camel.quarkus</groupId>
    <artifactId>camel-quarkus-docling</artifactId>
    <version>4.20.0 (CEQ 3.36.0)</version>
</dependency>

Endpoint Properties

Name	Type	Required	Default	Description
`operationId`	string	✓		The operation identifier
`contentInBody`	boolean		`false`	Include the content of the output file in the exchange body and delete the output file
`doclingServeUrl`	string		`http://localhost:5001`	Docling-serve API URL (e.g., http://localhost:5001)
`enableOCR`	boolean		`true`	Enable OCR processing for scanned documents
`includeLayoutInfo`	boolean		`false`	Show layout information with bounding boxes
`ocrLanguage`	string		`en`	Language code for OCR processing
`operation`	enum	✓	`CONVERT_TO_MARKDOWN`	The operation to perform
`outputFormat`	string		`markdown`	Output format for document conversion
`useDoclingServe`	boolean		`false`	Use docling-serve API instead of CLI command
`lazyStartProducer`	boolean		`false`	Whether the producer should be started lazy (on the first message). By starting lazy you can use this to allow CamelContext and routes to startup in situations where a producer may otherwise fail during starting and cause the route to fail being started. By deferring this startup to be lazy then the startup failure can be handled during routing messages via Camel's routing error handlers. Beware that when the first message is processed then creating and starting the producer may take a little time and prolong the total processing time of the processing.
`abortOnError`	boolean		`false`	Abort processing on error
`asyncPollInterval`	integer		`2000`	Polling interval for async conversion status in milliseconds
`asyncTimeout`	integer		`300000`	Maximum time to wait for async conversion completion in milliseconds
`doclingCommand`	string			Path to Docling Python executable or command
`doCodeEnrichment`	boolean		`false`	Enable code enrichment in document processing
`documentTimeout`	integer			Document processing timeout in seconds
`doFormulaEnrichment`	boolean		`false`	Enable formula enrichment in document processing
`doOcr`	boolean		`false`	Enable OCR processing in docling-serve API mode. When not set, the server uses its own defaults. Set enableOCR to false to explicitly disable OCR.
`doPictureClassification`	boolean		`false`	Enable picture classification in document processing
`doPictureDescription`	boolean		`false`	Enable picture description generation in document processing
`doTableStructure`	boolean		`false`	Enable table structure recognition
`forceOcr`	boolean		`false`	Force OCR processing even for digital documents
`imageExportMode`	string			Image export mode for referenced images
`imagesScale`	number			Scale factor for exported images
`includeImages`	boolean		`false`	Include images in the conversion output
`mdPageBreakPlaceholder`	string			Placeholder string for page breaks in markdown output
`ocrEngine`	string			OCR engine to use
`pdfBackend`	string			PDF parsing backend
`pipeline`	string			Processing pipeline to use
`processTimeout`	integer		`30000`	Timeout for Docling process execution in milliseconds
`tableCellMatching`	boolean		`false`	Enable table cell matching post-processing
`tableMode`	string			Table structure recognition mode
`useAsyncMode`	boolean		`false`	Use asynchronous conversion mode (docling-serve API only)
`workingDirectory`	string			Working directory for Docling execution
`batchFailOnFirstError`	boolean		`true`	Fail entire batch on first error (true) or continue processing remaining documents (false)
`batchParallelism`	integer		`4`	Number of parallel threads for batch processing
`batchSize`	integer		`10`	Number of documents to submit per sub-batch. Documents are partitioned into sub-batches of this size and each sub-batch is processed before starting the next one. Within each sub-batch, up to batchParallelism threads are used concurrently. This controls memory usage and back-pressure when processing large document sets.
`batchTimeout`	integer		`300000`	Maximum time to wait for batch completion in milliseconds
`splitBatchResults`	boolean		`false`	Split batch results into individual exchanges (one per document) instead of single BatchProcessingResults
`chunkingIncludeRawText`	boolean		`false`	Include raw text in chunk output
`chunkingMaxTokens`	integer			Maximum number of tokens per chunk for hybrid chunking
`chunkingMergePeers`	boolean		`true`	Whether to merge peer chunks in hybrid chunking
`chunkingTokenizer`	string			Tokenizer model for hybrid chunking (e.g. sentence-transformers/all-MiniLM-L6-v2)
`chunkingUseMarkdownTables`	boolean		`false`	Use markdown format for tables in chunk output
`includeMetadataInHeaders`	boolean		`true`	Include metadata in message headers when extracting metadata
`includeRawMetadata`	boolean		`false`	Include raw metadata as returned by the parser
`apiKeyHeader`	string		`X-API-Key`	Header name for API key authentication
`authenticationScheme`	enum		`NONE`	Authentication scheme (BEARER, API_KEY, NONE)
`authenticationToken`	string			Authentication token for docling-serve API (Bearer token or API key)
`maxFileSize`	integer		`52428800`	Maximum file size in bytes for processing
`oauthProfile`	string			OAuth profile name for obtaining an access token via the OAuth 2.0 Client Credentials grant. When set, the token is acquired from the configured identity provider and used as authenticationToken. Requires camel-oauth on the classpath.