> ## Documentation Index > Fetch the complete documentation index at: https://docs.reducto.ai/llms.txt > Use this file to discover all available pages before exploring further. > Reducto Extract API pulls structured fields from any document using a schema you define or schemaless configuration. # Extract ## OpenAPI ````yaml openapi.json post /extract openapi: 3.1.0 info: title: Reducto API version: v1.12.12-114-g9808b824f237 servers: - url: https://platform.reducto.ai security: [] paths: /extract: post: summary: Extract operationId: extract_extract_post requestBody: content: application/json: schema: oneOf: - $ref: '#/components/schemas/SyncExtractConfig' - $ref: '#/components/schemas/AsyncExtractConfig' required: true responses: '200': description: Successful Response content: application/json: schema: oneOf: - $ref: '#/components/schemas/V3ExtractResponse' - $ref: '#/components/schemas/AsyncExtractResponse' '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' security: - SkippableHTTPBearer: [] components: schemas: SyncExtractConfig: properties: input: anyOf: - type: string - items: type: string type: array - $ref: '#/components/schemas/UploadResponse' title: Input description: >- For parse/split/extract pipelines, the URL of the document to be processed. You can provide one of the following: 1. A publicly available URL 2. A presigned S3 URL 3. A reducto:// prefixed URL obtained from the /upload endpoint after directly uploading a document 4. A jobid:// prefixed URL obtained from a previous /parse invocation 5. A list of URLs (for multi-document pipelines, V3 API only) For edit pipelines, this should be a string containing the edit instructions parsing: $ref: '#/components/schemas/ParseOptions' description: >- The configuration options for parsing the document. If you are passing in a jobid:// URL for the file, then this configuration will be ignored. default: enhance: advanced_chart_agent: false agentic: [] intelligent_ordering: false summarize_figures: true retrieval: chunking: chunk_mode: disabled chunk_overlap: 0 embedding_optimized: false filter_blocks: [] formatting: add_page_markers: false include: [] merge_tables: false table_output_format: dynamic spreadsheet: clustering: accurate exclude: [] include: [] split_large_tables: enabled: true size: 50 settings: embed_pdf_metadata: false embed_pdf_metadata_dpi: 100 extract_document_properties: false extraction_mode: hybrid force_url_result: false hybrid_vpc: {} ocr_system: standard persist_results: false return_images: [] return_ocr_data: false instructions: $ref: '#/components/schemas/Instructions' description: The instructions to use for the extraction. default: schema: {} system_prompt: Be precise and thorough. settings: $ref: '#/components/schemas/ExtractSettings' description: The settings to use for the extraction. default: include_images: false optimize_for_latency: false force_url_result: false array_extract: false deep_extract: false citations: enabled: false numerical_confidence: true parent_block: full type: object required: - input title: SyncExtractConfig AsyncExtractConfig: properties: async: $ref: '#/components/schemas/config__v3__AsyncConfig' description: >- The configuration options for asynchronous processing (default synchronous). default: priority: false input: anyOf: - type: string - items: type: string type: array - $ref: '#/components/schemas/UploadResponse' title: Input description: >- For parse/split/extract pipelines, the URL of the document to be processed. You can provide one of the following: 1. A publicly available URL 2. A presigned S3 URL 3. A reducto:// prefixed URL obtained from the /upload endpoint after directly uploading a document 4. A jobid:// prefixed URL obtained from a previous /parse invocation 5. A list of URLs (for multi-document pipelines, V3 API only) For edit pipelines, this should be a string containing the edit instructions parsing: $ref: '#/components/schemas/ParseOptions' description: >- The configuration options for parsing the document. If you are passing in a jobid:// URL for the file, then this configuration will be ignored. default: enhance: advanced_chart_agent: false agentic: [] intelligent_ordering: false summarize_figures: true retrieval: chunking: chunk_mode: disabled chunk_overlap: 0 embedding_optimized: false filter_blocks: [] formatting: add_page_markers: false include: [] merge_tables: false table_output_format: dynamic spreadsheet: clustering: accurate exclude: [] include: [] split_large_tables: enabled: true size: 50 settings: embed_pdf_metadata: false embed_pdf_metadata_dpi: 100 extract_document_properties: false extraction_mode: hybrid force_url_result: false hybrid_vpc: {} ocr_system: standard persist_results: false return_images: [] return_ocr_data: false instructions: $ref: '#/components/schemas/Instructions' description: The instructions to use for the extraction. default: schema: {} system_prompt: Be precise and thorough. settings: $ref: '#/components/schemas/ExtractSettings' description: The settings to use for the extraction. default: include_images: false optimize_for_latency: false force_url_result: false array_extract: false deep_extract: false citations: enabled: false numerical_confidence: true parent_block: full queue_priority: $ref: '#/components/schemas/QueuePriority' description: >- Queue priority. 'batch' places the job in a lower-priority queue for non-urgent bulk work. 'auto' (alias: 'standard') uses the default queue. default: auto type: object required: - input title: AsyncExtractConfig V3ExtractResponse: properties: response_type: type: string const: v3_extract title: Response Type default: v3_extract job_id: anyOf: - type: string - type: 'null' title: Job Id usage: $ref: '#/components/schemas/ExtractUsage' studio_link: anyOf: - type: string - type: 'null' title: Studio Link description: The link to the studio pipeline for the document. confidence: anyOf: - type: string enum: - high - low - type: 'null' title: Confidence description: Optional document-level deep extract confidence label. confidence_reason: anyOf: - type: string - type: 'null' title: Confidence Reason description: Optional explanation for the document-level confidence label. result: anyOf: - {} - items: {} type: array title: Result description: >- The extracted response in your provided schema. This is a list of dictionaries. If disable_chunking is True (default), then it will be a list of length one. type: object required: - usage - result title: V3ExtractResponse AsyncExtractResponse: properties: job_id: type: string title: Job Id type: object required: - job_id title: AsyncExtractResponse HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError UploadResponse: properties: file_id: type: string title: File Id presigned_url: anyOf: - type: string - type: 'null' title: Presigned Url type: object required: - file_id title: UploadResponse ParseOptions: properties: enhance: $ref: '#/components/schemas/Enhance' default: agentic: [] advanced_chart_agent: false summarize_figures: true intelligent_ordering: false retrieval: $ref: '#/components/schemas/Retrieval' default: chunking: chunk_mode: disabled chunk_overlap: 0 filter_blocks: [] embedding_optimized: false formatting: $ref: '#/components/schemas/Formatting' default: add_page_markers: false table_output_format: dynamic merge_tables: false include: [] spreadsheet: $ref: '#/components/schemas/Spreadsheet' default: split_large_tables: enabled: true size: 50 include: [] clustering: accurate exclude: [] settings: $ref: '#/components/schemas/Settings' default: ocr_system: standard extraction_mode: hybrid force_url_result: false return_ocr_data: false return_images: [] embed_pdf_metadata: false extract_document_properties: false embed_pdf_metadata_dpi: 100 persist_results: false hybrid_vpc: {} type: object title: ParseOptions Instructions: properties: schema: title: Schema description: The JSON schema to use for the extraction. default: {} system_prompt: type: string title: System Prompt description: The system prompt to use for the extraction. default: Be precise and thorough. type: object title: Instructions ExtractSettings: properties: include_images: type: boolean title: Include Images description: If True, include images in the extraction. default: false optimize_for_latency: type: boolean title: Optimize For Latency description: >- If True, jobs will be processed with a higher throughput and priority at a higher cost. Defaults to False. default: false force_url_result: type: boolean title: Force Url Result description: Force the endpoint result to be returned in URL form. default: false array_extract: type: boolean title: Array Extract description: >- Deprecated: prefer deep_extract, which supersedes array extraction for complex and long (array-heavy) extractions via an agentic loop (at higher cost and latency). If True, use array extraction. default: false deep_extract: type: boolean title: Deep Extract description: >- If True, use Deep Extract, an agentic extraction mode that iteratively refines its output to achieve near-perfect accuracy. Best for complex documents where accuracy is critical. default: false citations: $ref: '#/components/schemas/Citations' description: The citations to use for the extraction. default: enabled: false numerical_confidence: true parent_block: full page_range: anyOf: - $ref: '#/components/schemas/PageRange' - items: $ref: '#/components/schemas/PageRange' type: array - items: type: integer type: array - items: type: string type: array - type: 'null' title: Page Range description: >- The page range to extract from (1-indexed). By default, the entire document is used. For spreadsheets, you can also provide a list of sheet names. type: object title: ExtractSettings config__v3__AsyncConfig: properties: metadata: title: Metadata description: JSON metadata included in webhook request body. Defaults to None. priority: type: boolean title: Priority description: >- Workers poll the priority queue ahead of the standard queue, so priority jobs start sooner when there is queued work; sync jobs are prioritized above async jobs by default. default: false webhook: anyOf: - $ref: '#/components/schemas/SvixWebhookConfig' - $ref: '#/components/schemas/DirectWebhookConfig' - type: 'null' title: Webhook description: The webhook configuration for the asynchronous processing. type: object title: AsyncConfig QueuePriority: type: string enum: - auto - standard - batch title: QueuePriority description: >- Customer-facing queue priority for parse jobs. ``AUTO`` and ``STANDARD`` are wire-level synonyms — both mean "default queue." Downstream code only branches on ``BATCH``, so the difference is purely lexical. Both values are first-class enum members so a rolling deploy is safe in either direction: a pre-#6134 pod (enum = ``{auto, batch}``) and a post-fix pod (enum = ``{auto, standard, batch}``) can both deserialise payloads the other produces. Migration plan (three stages, each a separate PR fully deployed before the next): 1. **Widen** *(this PR)*: enum = ``{AUTO, STANDARD, BATCH}``, ``default=AUTO``. Compatible with the still-running pre-#6134 prod pods after yesterday's rollback. 2. **Flip default**: same enum, ``default=STANDARD``. Ship after stage 1 is everywhere. 3. **Narrow** *(future, optional)*: drop ``AUTO`` once every Redis payload and SDK has migrated. Until then it stays as an accepted alias. ExtractUsage: properties: num_pages: type: integer title: Num Pages num_fields: type: integer title: Num Fields credits: anyOf: - type: number - type: 'null' title: Credits extract_mode: anyOf: - type: string enum: - super_agent - extract - spreadsheet_agent - type: 'null' title: Extract Mode usage_breakdown: anyOf: - $ref: '#/components/schemas/ExtractUsageBreakdown' - type: 'null' description: >- Raw usage quantities. Only set for accounts on the new pricing model; credit fields are omitted for those accounts. type: object required: - num_pages - num_fields title: ExtractUsage ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type input: title: Input ctx: type: object title: Context type: object required: - loc - msg - type title: ValidationError Enhance: properties: agentic: items: anyOf: - $ref: '#/components/schemas/TableAgentic' - $ref: '#/components/schemas/FigureAgentic' - $ref: '#/components/schemas/TextAgentic' type: array title: Agentic description: >- For legacy Parse, agentic processing uses vision language models to improve text, table, or figure extraction. With r-1, use agentic processing for custom prompts or advanced chart extraction. Agentic processing adds latency. default: [] advanced_chart_agent: type: boolean title: Advanced Chart Agent description: >- If True, run advanced chart extraction on figures classified as charts, without requiring a figure-scoped agentic entry. Returns full structured series data (chart_data) plus a reconstruction image re-drawn from that data. Higher latency. Defaults to False. default: false summarize_figures: type: boolean title: Summarize Figures description: >- For legacy Parse, if True, summarize figures using a separate vision language model. r-1 generates figure descriptions natively and ignores this setting. Defaults to True. default: true intelligent_ordering: type: boolean title: Intelligent Ordering description: >- For legacy Parse, if True, use an advanced vision language model to improve reading order accuracy, with a small increase in latency. r-1 handles reading order natively and ignores this setting. Defaults to False. default: false type: object title: Enhance Retrieval: properties: chunking: $ref: '#/components/schemas/Chunking' default: chunk_mode: disabled chunk_overlap: 0 filter_blocks: items: type: string enum: - Header - Footer - Title - Section Header - Page Number - List Item - Figure - Table - Key Value - Text - Comment - Signature type: array title: Filter Blocks description: >- A list of block types to filter out from 'content' and 'embed' fields. By default, no blocks are filtered. default: [] embedding_optimized: type: boolean title: Embedding Optimized description: If True, use embedding optimized mode. Defaults to False. default: false type: object title: Retrieval Formatting: properties: add_page_markers: type: boolean title: Add Page Markers description: >- If True, add page markers to the output. Defaults to False. Useful for extracting data with page specific information. default: false table_output_format: type: string enum: - html - json - md - jsonbbox - dynamic - csv title: Table Output Format description: >- The table output format. Defaults to dynamic, which returns md for simpler tables and html for more complex tables. r-1 does not support jsonbbox. default: dynamic merge_tables: type: boolean title: Merge Tables description: >- A flag to indicate if consecutive tables with the same number of columns should be merged. Defaults to False. default: false include: items: type: string enum: - change_tracking - highlight - comments - hyperlinks - signatures - ignore_watermarks type: array title: Include description: >- For legacy Parse, the formatting details to include in the output. r-1 handles highlights, signatures, and watermarks natively and ignores those values. default: [] type: object title: Formatting Spreadsheet: properties: split_large_tables: $ref: '#/components/schemas/SplitLargeTables' default: enabled: true size: 50 include: items: type: string enum: - cell_colors - formula - dropdowns type: array title: Include description: >- Whether to include cell color, formula, and dropdown information in the output. default: [] clustering: type: string enum: - accurate - fast - disabled title: Clustering description: >- In a spreadsheet with different tables inside, we enable splitting up the tables by default. Accurate mode applies more powerful models for superior accuracy, at 5× the default per-cell rate. Disabling will register as one large table. default: accurate exclude: items: type: string enum: - hidden_sheets - hidden_rows - hidden_cols - styling - spreadsheet_images type: array title: Exclude description: Whether to exclude hidden sheets, rows, or columns in the output. default: [] max_cell_count: anyOf: - type: integer minimum: 1 - type: 'null' title: Max Cell Count description: >- Maximum total non-empty cells allowed across all sheets. If exceeded, the request is rejected with a 422 error. Set to null to disable the limit. Defaults to null. type: object title: Spreadsheet Settings: properties: ocr_system: type: string enum: - standard - legacy title: Ocr System description: >- The OCR system for legacy Parse. Standard is the best multilingual OCR system. Legacy supports Germanic languages and remains available for backwards compatibility. r-1 uses native full-page processing and ignores this setting. Defaults to standard. default: standard extraction_mode: type: string enum: - ocr - hybrid title: Extraction Mode description: >- The text extraction method for legacy Parse. OCR uses optical character recognition only. Hybrid combines OCR with embedded PDF text. r-1 uses native full-page processing and ignores this setting. Defaults to hybrid. default: hybrid force_url_result: type: boolean title: Force Url Result description: Force the result to be returned in URL form. default: false force_file_extension: anyOf: - type: string - type: 'null' title: Force File Extension description: >- Force the URL to be downloaded as a specific file extension (e.g. `.png`). return_ocr_data: type: boolean title: Return Ocr Data description: If True, return OCR data in the result. Defaults to False. default: false return_images: items: type: string enum: - figure - table - page type: array title: Return Images description: >- Whether to return images for the specified block types. 'page' returns full page images. By default, no images are returned. default: [] embed_pdf_metadata: type: boolean title: Embed Pdf Metadata description: >- If True, embed OCR metadata into the returned PDF. Defaults to False. default: false extract_document_properties: type: boolean title: Extract Document Properties description: >- If True, return properties embedded in the original document. Defaults to False. default: false embed_pdf_metadata_dpi: type: integer maximum: 250 minimum: 50 title: Embed Pdf Metadata Dpi description: >- Render DPI used when rasterizing the source PDF before embedding the OCR text layer (only applies when ``embed_pdf_metadata`` is True). Lower values produce dramatically smaller output PDFs; higher values preserve more detail when zoomed past 200%. Defaults to 100 (good for on-screen viewing); raise toward the source scan DPI for crisper output. Min 50, max 250. default: 100 persist_results: type: boolean title: Persist Results description: If True, persist the results indefinitely. Defaults to False. default: false tenant_throttling: anyOf: - $ref: '#/components/schemas/TenantThrottling' - type: 'null' description: >- Per-tenant throttling for multi-tenant applications. Tag each request with your tenant's id to bound how much of your account's concurrency a single tenant can consume. Account-level throttles still apply. timeout: anyOf: - type: number - type: 'null' title: Timeout description: The timeout for the job in seconds. page_range: anyOf: - $ref: '#/components/schemas/PageRange' - items: $ref: '#/components/schemas/PageRange' type: array - items: type: integer type: array - items: type: string type: array - type: 'null' title: Page Range description: >- The page range to process (1-indexed). By default, the entire document is processed. For spreadsheets, you can also provide a list of sheet names. document_password: anyOf: - type: string - type: 'null' title: Document Password description: Password to decrypt password-protected documents. hybrid_vpc: $ref: '#/components/schemas/HybridVpcSettings' description: Hybrid VPC request-scoped settings. default: {} model: anyOf: - type: string enum: - r-1 - legacy - type: 'null' title: Model description: >- The parse model to use. 'r-1' is the R-1 full-page parse model, which parses each page in a single generation. 'legacy' is the previous parsing pipeline. Defaults to 'legacy' unless your organization was created on the r-1 plan, in which case it defaults to 'r-1'. type: object title: Settings Citations: properties: enabled: type: boolean title: Enabled description: If True, include citations in the extraction. default: false numerical_confidence: type: boolean title: Numerical Confidence description: >- If True, enable numeric citation confidence scores. Defaults to True. default: true parent_block: $ref: '#/components/schemas/CitationParentBlockMode' description: >- How much of the source parse block to embed on each citation's parentBlock. 'full' (default) embeds the verbatim source-block HTML in parentBlock.content. 'bbox_only' suppresses parentBlock.content (returned as an empty string) while keeping parentBlock.bbox and all citation-level fields — this can drastically shrink responses on table-heavy schemas where the same source block is cited many times. default: full type: object title: Citations PageRange: properties: start: anyOf: - type: integer - type: 'null' title: Start description: The page number to start processing from (1-indexed). end: anyOf: - type: integer - type: 'null' title: End description: The page number to stop processing at (1-indexed). type: object title: PageRange SvixWebhookConfig: properties: mode: type: string const: svix title: Mode default: svix channels: items: type: string type: array title: Channels description: >- A list of Svix channels the message will be delivered down, omit to send to all channels. type: object title: SvixWebhookConfig DirectWebhookConfig: properties: mode: type: string const: direct title: Mode default: direct url: type: string title: Url type: object required: - url title: DirectWebhookConfig ExtractUsageBreakdown: properties: extract_model: type: string enum: - Extract - Deep Extract title: Extract Model tier: type: string enum: - Default - Batch title: Tier default: Default extract_pages: type: integer title: Extract Pages default: 0 extract_fields: type: integer title: Extract Fields default: 0 ocr_pages: type: integer title: Ocr Pages default: 0 charts: type: integer title: Charts default: 0 prompted_blocks: type: integer title: Prompted Blocks default: 0 type: object required: - extract_model title: ExtractUsageBreakdown description: >- Raw extract quantities for accounts on the new pricing model. ``extract_fields`` is reported but not billed at launch. The add-on quantities (``ocr_pages``, ``charts``, ``prompted_blocks``) come from the parse bundled into the extract job; its page cost is covered by ``extract_pages`` but its add-ons are billed separately. ``tier`` is "Batch" when the job ran on the batch queue, which takes the batch discount on the rate card. TableAgentic: properties: scope: type: string const: table title: Scope prompt: anyOf: - type: string - type: 'null' title: Prompt description: Custom prompt for table agentic. mode: type: string enum: - default - auto - max title: Mode description: >- Mode for table agentic: 'default' selectively applies enrichment only to tables likely to benefit, and 'max' runs enrichment on all tables. default: default type: object required: - scope title: TableAgentic FigureAgentic: properties: scope: type: string const: figure title: Scope prompt: anyOf: - type: string - type: 'null' title: Prompt description: Custom prompt for figure agentic. advanced_chart_agent: type: boolean title: Advanced Chart Agent description: >- If True, run advanced chart extraction on figures classified as charts: an agentic extractor that returns full structured series data (chart_data) plus a reconstruction image re-drawn from that data (extra.chart_reconstruction). Higher latency. Defaults to False. default: false return_overlays: type: boolean title: Return Overlays description: >- If True, return overlays for the figure. This is so you can use the overlays to double check the quality of the extraction default: false type: object required: - scope title: FigureAgentic TextAgentic: properties: scope: type: string const: text title: Scope prompt: anyOf: - type: string - type: 'null' title: Prompt description: >- Custom instructions for agentic text. Note: This only applies to form regions (key-value). type: object required: - scope title: TextAgentic Chunking: properties: chunk_mode: type: string enum: - variable - section - page - disabled - block - page_sections title: Chunk Mode description: >- Choose how to partition chunks. Variable mode chunks by character length and visual context. Section mode chunks by section headers. Page mode chunks according to pages. Page sections mode chunks first by page, then by sections within each page. Disabled returns one single chunk. default: disabled chunk_size: anyOf: - type: integer - type: 'null' title: Chunk Size description: >- The approximate size of chunks (in characters) that the document will be split into. Defaults to null, in which case the chunk size is variable between 250 - 1500 characters. chunk_overlap: type: integer title: Chunk Overlap description: >- Number of characters of overlap to include from adjacent chunks. Defaults to 0. default: 0 type: object title: Chunking SplitLargeTables: properties: enabled: type: boolean title: Enabled description: If True, split large tables into smaller tables. Defaults to True. default: true size: anyOf: - type: integer - $ref: '#/components/schemas/SplitLargeTableSizes' title: Size description: >- The size of the tables to split into. Defaults to 50. Use 'row' and 'column' to independently specify the number of rows and columns to include when splitting. If you only want to split by rows or columns, set the other value to None. default: 50 type: object title: SplitLargeTables TenantThrottling: properties: tenant_id: type: string maxLength: 256 minLength: 1 title: Tenant Id description: >- Your identifier for the tenant (customer, workspace, organization) this request belongs to. Used only for noisy-neighbor throttling inside your account. max_share: type: number maximum: 1 exclusiveMinimum: 0 title: Max Share description: >- Maximum fraction of your account's concurrency ceiling this tenant may use, between 0 (exclusive) and 1. Defaults to 0.5. default: 0.5 type: object required: - tenant_id title: TenantThrottling HybridVpcSettings: properties: environment: anyOf: - type: string - type: 'null' title: Environment description: >- Named Hybrid VPC environment to use for this request. Only applies when your organization has Hybrid VPC environments configured. type: object title: HybridVpcSettings CitationParentBlockMode: type: string enum: - full - bbox_only title: CitationParentBlockMode description: >- How much of the source parse block to embed on each citation's ``parentBlock``. ``FULL`` embeds the verbatim source-block HTML in ``parentBlock.content``. ``BBOX_ONLY`` returns ``parentBlock.content`` as an empty string while keeping ``parentBlock.bbox`` and all citation-level fields — on table-heavy schemas the same source block is re-embedded into every citation that points into it, so the content can dominate the response (90%+ of bytes) while being redundant with the parse output. SplitLargeTableSizes: properties: row: anyOf: - type: integer - type: 'null' title: Row description: >- The number of rows to include in each chunk when splitting large tables. Does not chunk rows if set to None. column: anyOf: - type: integer - type: 'null' title: Column description: >- The number of columns to include in each chunk when splitting large tables. Does not chunk columns if set to None. type: object title: SplitLargeTableSizes securitySchemes: SkippableHTTPBearer: type: http scheme: bearer ````