> ## Documentation Index > Fetch the complete documentation index at: https://docs.meibel.ai/llms.txt > Use this file to discover all available pages before exploring further. # Get structured parse result > Download the fully structured parse result (the json format): pages, typed elements, tables, chart data, chart OCR text, and bounding boxes. The response schema (StructuredDocument) is defined by the parsing engine and hoisted into this spec by the OpenAPI generator. ## OpenAPI ````yaml /api-v2.json get /documents/{job_id}/structured openapi: 3.1.0 info: title: Meibel AI API summary: Meibel Gateway Service description: >- Our API allows you to interact with our services. Read the [docs](https://docs.meibel.ai) to learn how to use it. version: 0.6.4 servers: - url: https://api.meibel.ai/v2 description: Meibel API (v2) - url: https://api.dev.meibel.ai/v2 description: Meibel Dev API (v2) - url: http://localhost:8000/v2 description: Local Development security: [] tags: - name: Datasources description: v2 datasource management - name: Data Elements description: v2 data element CRUD and search - name: File Upload description: Upload files to a datasource and track upload progress - name: Ingest description: Trigger and monitor datasource ingestion - name: Datasource Downloads description: Export all datasource content as a zip archive - name: Table Descriptions description: v2 table and column descriptions - name: Metadata Model Catalog description: v2 metadata model catalog - name: Confidence Scoring description: v2 confidence scoring jobs and summaries - name: Documents description: Parse and transform documents into structured data - name: Agents description: Agent definition management - name: Sessions description: Agent session (execution) management and chat - name: Execution Policies description: Data access and tool usage constraints scoped to agent sessions - name: Artifact Schemas description: Agent artifact schema management paths: /documents/{job_id}/structured: get: tags: - Documents summary: Get structured parse result description: >- Download the fully structured parse result (the json format): pages, typed elements, tables, chart data, chart OCR text, and bounding boxes. The response schema (StructuredDocument) is defined by the parsing engine and hoisted into this spec by the OpenAPI generator. operationId: getStructuredDocumentResult parameters: - name: job_id in: path required: true schema: type: string title: Job Id responses: '200': description: Successful Response content: application/json: schema: $ref: '#/components/schemas/ParseStructuredDocument' '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' security: - APIKeyHeader: [] x-codeSamples: - lang: Python label: Python source: |- from meibel import MeibelClient client = MeibelClient(api_key="your-api-key") result = client.documents.get_structured_result("job_id_value") print(result) - lang: TypeScript label: TypeScript source: >- import { MeibelClient } from 'meibel'; const client = new MeibelClient({ apiKey: 'your-api-key' }); const result = await client.documents.getStructuredResult('job_id_value'); console.log(result); - lang: Go label: Go source: >- import v2 "github.com/meibel-ai/meibel-go/v2" client := v2.NewClient(v2.WithAPIKey("your-api-key")) ctx := context.Background() result, err := client.Documents.GetStructuredResult(ctx, "job_id_value") if err != nil { log.Fatal(err) } fmt.Println(result) components: schemas: ParseStructuredDocument: title: StructuredDocument description: 'The final output of the pipeline: a fully structured document.' type: object required: - confidence - num_pages - pages properties: confidence: description: Aggregate confidence scores across all pages. allOf: - $ref: '#/components/schemas/ParseConfidenceScores' format: description: >- Detected input format (e.g. "pdf", "docx", "markdown"). `None` for documents processed before this field was added. type: - string - 'null' gpu_ms: description: >- Total GPU inference time in milliseconds across all stages (layout detection + table encoder + table decoder + OCR). Zero for non-PDF formats that don't use the ML pipeline. default: 0 type: integer format: uint64 minimum: 0 num_pages: description: Number of pages in the source document. type: integer format: uint minimum: 0 ocr_pages: description: Number of pages that required OCR (had no extractable text). default: 0 type: integer format: uint32 minimum: 0 orientation_pages: description: >- Number of pages where orientation detection applied a non-zero rotation. default: 0 type: integer format: uint32 minimum: 0 pages: description: Per-page structured content. type: array items: $ref: '#/components/schemas/ParseStructuredPage' remote_regions: description: >- Number of regions dispatched to remote endpoints (charts, formulas, seals). default: 0 type: integer format: uint32 minimum: 0 HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError ParseConfidenceScores: description: Aggregate confidence scores for the document. type: object required: - mean_layout_confidence - min_layout_confidence - num_elements - num_tables properties: mean_layout_confidence: description: Mean layout detection confidence across all elements. type: number format: double min_layout_confidence: description: Minimum layout detection confidence across all elements. type: number format: double num_elements: description: Total number of layout elements detected. type: integer format: uint minimum: 0 num_tables: description: Number of tables recognized. type: integer format: uint minimum: 0 ParseStructuredPage: description: Structured content for a single page. type: object required: - elements - page_bbox - page_number properties: elements: description: Document elements in reading order. type: array items: $ref: '#/components/schemas/ParseDocumentElement' extracted_page_no: description: >- The page number printed on the mini-page itself ("Page 62"): read off the page by the grid detector when legible, inferred from the confirmed reading order when obscured. type: - integer - 'null' format: uint32 minimum: 0 image_size: description: Image dimensions in pixels (if a page image was provided). type: - array - 'null' items: - type: integer format: uint32 minimum: 0 - type: integer format: uint32 minimum: 0 maxItems: 2 minItems: 2 ocr_applied: description: Whether OCR was applied to this page. default: false type: boolean ocr_dpi: description: >- Adaptive OCR rendering DPI used for this page (150-600, computed from page size). Only present when OCR was applied. type: - integer - 'null' format: uint32 minimum: 0 ocr_reason: description: >- OCR detection signal breakdown (e.g. "enc=0.85(majority_bad), font=0.15(rich_fonts), ..."). type: - string - 'null' ocr_score: description: >- OCR detection composite score (0.0 = definitely has text, 1.0 = definitely needs OCR). Present even when OCR was not applied. type: - number - 'null' format: double orientation_degrees: description: >- Orientation correction applied to this page in degrees (0, 90, 180, 270). 0 means no correction was needed or orientation detection was disabled. default: 0 type: integer format: uint16 minimum: 0 page_bbox: description: Page dimensions in PDF points (bottom-left origin). allOf: - $ref: '#/components/schemas/ParseBBox' page_number: description: Page number (0-indexed). type: integer format: uint minimum: 0 physical_page: description: 0-based index of the physical PDF page this output page came from. type: - integer - 'null' format: uint minimum: 0 quadrant: description: >- Which grid cell of the physical page this output page is, numbered in reading order. type: - integer - 'null' format: uint8 minimum: 0 sheet_rect: description: >- Rectangle this output page occupies on the physical page, in the physical page's coordinate space. The viewer uses this to reassemble split pages onto the source PDF. anyOf: - $ref: '#/components/schemas/ParseBBox' - type: 'null' transcript_lines: description: >- Present when the page was detected as a line-numbered legal transcript. The gutter numbers are stripped from body text and kept here so page:line citations remain resolvable. type: - array - 'null' items: $ref: '#/components/schemas/ParseTranscriptLine' 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 ParseDocumentElement: description: >- A single document element (text block, table, figure, etc.) in reading order. type: object required: - bbox - confidence - label - reading_order - text properties: bbox: description: >- Bounding box in pixel coordinates (top-left origin). Normalized to [0,1] for annotated export. allOf: - $ref: '#/components/schemas/ParseBBox' chart_data: description: If this is a Chart element, the digitized plot data. anyOf: - $ref: '#/components/schemas/ParseChartData' - type: 'null' confidence: description: Confidence score from the layout model. type: number format: double heading_level: description: >- Heading level (1-6) for Title/SectionHeader elements. Determined by font-size clustering: largest font → H1, decreasing → H2-H6. `None` for non-heading elements. type: - integer - 'null' format: uint8 minimum: 0 label: description: Layout label from the detection model. allOf: - $ref: '#/components/schemas/ParseLayoutLabel' ocr_text: description: >- Recognized text on a Chart region, with positions and provenance. Sibling to `chart_data` so it survives when chart_data is None. type: array items: $ref: '#/components/schemas/ParseChartText' reading_order: description: Position in reading order (0-indexed within page). type: integer format: uint minimum: 0 table: description: If this is a Table element, the recognized table structure. anyOf: - $ref: '#/components/schemas/ParseTable' - type: 'null' text: description: Text content (assembled from cells within this region). type: string ParseBBox: description: Axis-aligned bounding box. type: object required: - x0 - x1 - y0 - y1 properties: x0: type: number format: double x1: type: number format: double y0: type: number format: double y1: type: number format: double ParseTranscriptLine: description: A transcript gutter line number stripped from the text flow. type: object required: - bbox - number properties: bbox: description: Bounding box of the number in pixel coordinates (top-left origin). allOf: - $ref: '#/components/schemas/ParseBBox' number: description: The printed line number (1-25/26). type: integer format: uint32 minimum: 0 ParseChartData: type: object required: - categories - chart_type - modality - overall_confidence - plot_area - series - warnings - x_axis properties: categories: items: type: string type: array chart_type: $ref: '#/components/schemas/ParseChartType' modality: $ref: '#/components/schemas/ParseModality' overall_confidence: type: number format: double plot_area: $ref: '#/components/schemas/ParseDualBBox' series: type: array items: $ref: '#/components/schemas/ParseSeries' title: type: - string - 'null' warnings: items: type: string type: array x_axis: $ref: '#/components/schemas/ParseAxisCalibration' y_axis_left: anyOf: - $ref: '#/components/schemas/ParseAxisCalibration' - type: 'null' y_axis_right: anyOf: - $ref: '#/components/schemas/ParseAxisCalibration' - type: 'null' ParseLayoutLabel: description: >- Layout element types for document layout detection. PP-DocLayoutV3 produces 25 classes (indices 0-24) which are mapped to these labels. Legacy heron labels (Checkbox*, Form, KeyValueRegion, ListItem) are retained for backward compatibility but no longer emitted by the model. oneOf: - description: Caption for a figure or table. type: string enum: - Caption - description: Chart or plot. type: string enum: - Chart - description: Footnote text. type: string enum: - Footnote - description: Mathematical formula (display or inline). type: string enum: - Formula - description: Bullet or numbered list item (legacy, not in V3). type: string enum: - ListItem - description: Footer/running foot at bottom of page. type: string enum: - PageFooter - description: Header/running head at top of page. type: string enum: - PageHeader - description: Figure, diagram, or image. type: string enum: - Picture - description: Seal or stamp. type: string enum: - Seal - description: Section heading. type: string enum: - SectionHeader - description: Table — triggers table structure recognition. type: string enum: - Table - description: Running body text. type: string enum: - Text - description: Document or section title. type: string enum: - Title - description: Table of contents / document index. type: string enum: - DocumentIndex - description: Code block or monospaced listing. type: string enum: - Code - description: Selected checkbox (legacy, not in V3). type: string enum: - CheckboxSelected - description: Unselected checkbox (legacy, not in V3). type: string enum: - CheckboxUnselected - description: Form region (legacy, not in V3). type: string enum: - Form - description: Key-value region (legacy, not in V3). type: string enum: - KeyValueRegion ParseChartText: description: >- One recognized text run on a chart, with dual-space position and provenance. type: object required: - bbox - confidence - source - text properties: bbox: $ref: '#/components/schemas/ParseDualBBox' confidence: type: number format: double source: $ref: '#/components/schemas/ParseChartTextSource' text: type: string ParseTable: description: >- A recognized table structure on a page. Produced by the table structure recognition model (TableFormer via ONNX). type: object required: - bbox - cells - num_cols - num_rows - page_number properties: bbox: description: Bounding box of the entire table. allOf: - $ref: '#/components/schemas/ParseBBox' cells: description: >- Table cells (not to be confused with TextCell — these are table grid cells). type: array items: $ref: '#/components/schemas/ParseTableCell' num_cols: description: Number of columns. type: integer format: uint minimum: 0 num_rows: description: Number of rows. type: integer format: uint minimum: 0 page_number: description: Page number this table belongs to (0-indexed). type: integer format: uint minimum: 0 ParseChartType: type: string enum: - Line - Scatter - Bar - Area - Pie - Mixed - Unknown ParseModality: type: string enum: - Vector - Raster ParseDualBBox: description: >- A bbox in both coordinate spaces. `pixel` = rendered-image space (top-left origin); `pdf` = page points (bottom-left origin, BOTTOMLEFT). type: object required: - pdf - pixel properties: pdf: $ref: '#/components/schemas/ParseBBox' pixel: $ref: '#/components/schemas/ParseBBox' ParseSeries: type: object required: - points - style - y_axis properties: color: type: - array - 'null' items: type: integer format: uint8 minimum: 0 maxItems: 3 minItems: 3 dash_pattern: description: >- PDF dash array identifying the series alongside color. `Some(vec![])` = solid. Distinguishes monochrome series (Apple 10-K case). `None` = unknown (raster). type: - array - 'null' items: type: number format: double name: type: - string - 'null' points: type: array items: $ref: '#/components/schemas/ParseDataPoint' style: $ref: '#/components/schemas/ParseSeriesStyle' y_axis: $ref: '#/components/schemas/ParseYAxisRef' ParseAxisCalibration: type: object required: - data_range - pixel_to_data - scale - ticks properties: data_range: type: array items: - type: number format: double - type: number format: double maxItems: 2 minItems: 2 pixel_to_data: $ref: '#/components/schemas/ParseAffineFit' scale: $ref: '#/components/schemas/ParseAxisScale' ticks: type: array items: $ref: '#/components/schemas/ParseTickMark' title: type: - string - 'null' unit: type: - string - 'null' ParseChartTextSource: description: Where a `ChartText` came from. oneOf: - description: >- Sliced from the parser's PDF text cells (vector charts). PDF cells carry no confidence, so a synthesized 1.0 is used. type: string enum: - PdfText - description: >- Produced by the det+rec OCR engine on the chart crop (raster charts). type: string enum: - Ocr ParseTableCell: description: A single cell in a recognized table. type: object required: - bbox - col - col_span - is_header - row - row_span - text properties: bbox: description: Bounding box of this table cell. allOf: - $ref: '#/components/schemas/ParseBBox' col: description: Column index (0-indexed). type: integer format: uint minimum: 0 col_span: description: Number of columns this cell spans (1 = no spanning). type: integer format: uint minimum: 0 is_header: description: Whether this is a header cell. type: boolean row: description: Row index (0-indexed). type: integer format: uint minimum: 0 row_span: description: Number of rows this cell spans (1 = no spanning). type: integer format: uint minimum: 0 text: description: >- Text content of this cell (assembled from text cells within the bbox). type: string ParseDataPoint: description: >- One digitized data value. `x` is in data units after inversion; when `x_is_category` it is the float index into `ChartData::categories`; for a DateTime axis it is a unix timestamp (seconds). type: object required: - bbox - confidence - source - x - x_is_category - 'y' properties: bbox: $ref: '#/components/schemas/ParseDualBBox' confidence: type: number format: double source: $ref: '#/components/schemas/ParseValueSource' x: type: number format: double x_is_category: type: boolean 'y': type: number format: double ParseSeriesStyle: type: string enum: - Line - Scatter - Bar - Area - PieSlice ParseYAxisRef: description: >- Which Y axis a series reads against. `Ambiguous` is the fallback defined canonically in the spec §7f (dual-Y detected ∧ color match inconclusive ∧ no VLM) — emitted with a warning rather than a wrong guess. type: string enum: - Left - Right - Ambiguous ParseAffineFit: description: >- Linear fit `value = slope * pixel_pos + intercept` (on log10(value) for Log10, on unix-timestamp for DateTime). `r_squared` retained to warn on weak fits. type: object required: - intercept - r_squared - slope properties: intercept: type: number format: double r_squared: type: number format: double slope: type: number format: double ParseAxisScale: type: string enum: - Linear - Log10 - Categorical - DateTime ParseTickMark: type: object required: - label_bbox - pixel_pos - source - value properties: label_bbox: $ref: '#/components/schemas/ParseDualBBox' pixel_pos: type: number format: double source: $ref: '#/components/schemas/ParseValueSource' value: type: number format: double ParseValueSource: description: Provenance for a digitized value — drives confidence and auditability. type: string enum: - VectorPath - RasterMask - RasterMarker - VlmAdjudicated securitySchemes: APIKeyHeader: type: apiKey in: header name: Meibel-API-Key ````