How Inline XBRL 1.1, ESEF ZIP, and XBRL 2.1 filings parse
Financial teams often need to check tagged statutory figures before filing annual accounts or reviewing published reports. This local iXBRL viewer inspects .xhtml, .html, .htm, .xml, and packaged .zip submissions up to 40 MB entirely inside a client-side Web Worker following the Inline XBRL Part 1 Specification 1.1.
| File format | Detected standard | Package priority | Parser behavior | Special handling |
|---|---|---|---|---|
.xhtml |
Inline XBRL 1.1 | Highest inside reports/ |
Extracts ix:nonFraction and ix:nonNumeric |
Keeps nested disclosure tags |
.html / .htm |
Inline XBRL 1.1 | Fallback after .xhtml |
Strips HTML styles and scripts | Resolves entities |
.xml |
XBRL 2.1 XML | Below Inline HTML files | Reads elements with contextRef |
Deprioritizes _cal, _pre, _lab, _def |
.zip |
ESEF / UKSEF bundle | Ranked automatically | Unpacks up to 80 MB total | Skips __MACOSX/ and reportPackage.xml |
| Sample button | Inline XBRL 1.1 | Synthetic IFRS demo | Loads harbor-esef-2025-sample.xhtml |
Includes scale="3" and dimensions |
When scanning Inline XBRL documents, balanced tag tracking pairs nested ix:nonNumeric disclosures cleanly instead of truncating early at inner closing tags. Narrative fragments linked through continuedAt attributes are joined across matching ix:continuation IDs using cycle protection, while <ix:exclude> blocks are omitted and <ix:hidden> facts remain marked.
Inside compressed ESEF or UKSEF packages, the scoring engine selects files matching a reports/ folder ending in .xhtml ahead of root HTML files, while deprioritizing linkbases and catalog files.
Audit reviewers moving from statutory annual statements down to underlying statutory general ledgers in France often follow up report inspection with our FEC file checker. Both workflows keep sensitive pre-filing figures off external networks while exposing structural anomalies early.
How are scaled numeric facts, signs, and contexts resolved?
Display text inside annual reports frequently shows rounded thousands or bracketed losses that break generic spreadsheet imports. Our parser strips currency symbols ($, £, €, ¥) and non-breaking spaces, honors xsi:nil="true" nulls, treats zero-dash tokens as 0, and evaluates formatting rules before computing arithmetic totals.
Scaling factors, decimals, negative signs, and ixt formats
Comma-decimal transformations matching commadecimal or numspacecomma strip thousands dots or spaces before turning commas into decimal points, whereas dotdecimal and numspacedot preserve dots. When format is omitted, separator positions determine the decimal mark automatically, matching the structure defined by the XBRL 2.1 Instance Specification.
Bracketed figures like (640) become negative unless sign="+" is explicitly set, and sign="-" inverts positive text values such as finance costs. The resulting number is multiplied by ten raised to scale (for instance scale="3" for thousands) and rounded to six decimal places to prevent binary float tails.
Zero-dash tokens such as zerodash, numdash, fixed-zero, nil, and plain hyphens evaluate cleanly to 0, while HTML non-breaking spaces and XML character entities normalize before display or numeric export.
Resolving entity contexts, period dates, and dimensions
Every fact resolves its contextRef against <xbrli:context> blocks to capture entity schemes, identifiers, instant dates, startDate to endDate durations, or forever periods. Explicit and typed segment or scenario dimensions such as ifrs-full:SegmentsAxis are concatenated into readable axis=member pairs alongside <xbrli:unit> ratios like ISO 4217 EUR or EUR/shares.
Controllers reconciling German subsidiary trial balances extracted from consolidated European reports can also convert DATEV booking CSV files inside the same browser suite. Using a consistent local pipeline avoids copy-paste rounding drift across reporting periods.
Searching tagged facts and exporting to Excel or filtered CSV
Once parsing completes, sidebar summary counters display exact totals for numeric facts, text disclosures, contexts, and units across the active document. Filter buttons toggle between All, Numeric, and Disclosures, while real-time search matches concept QNames, labels, scaled numbers, raw text, periods, units, and dimensions.
Filtering numeric vs text disclosures and 300-row previews
Browser tables display the first 300 matching facts and shorten narrative disclosures to 180 characters onscreen so multi-thousand-tag filings stay responsive. Electronic reporting packages submitted under the ESMA European Single Electronic Format portal often exceed several thousand facts, so downloads always export the complete filtered dataset.
Four-sheet Excel workbooks and RFC 4180 CSV downloads
Selecting Excel workbook (.xlsx) generates true numeric cells rather than formatted strings, whereas Filtered facts CSV (.csv) writes RFC 4180 comma-delimited rows with a UTF-8 BOM across 14 structured columns. All text exports run through sanitizeCsvTable to neutralize leading formula characters (=, +, -, @, |).
- Numeric Facts sheet: Stores concept QName, prefix, readable label, native numeric
Scaled Value, originalRaw Text, unit, scale exponent, decimals, period, entity, dimensions, and hidden status. - Text Disclosures sheet: Captures long
ix:nonNumericaccounting policies and notes up to 10,000 characters per fact alongside context and entity metadata. - Contexts and Validation sheets: Lists every parsed context period with dimension strings plus all structural errors and warnings detected during the pass.
CSV downloads write 14 columns covering Concept, Prefix, Label, Kind, Scaled Value, Raw Text, Unit, Scale, Decimals, Period, Entity, Dimensions, Context ID, and Hidden flags.
Finance teams verifying UK statutory filings alongside supplier payment runs can pair these exported spreadsheets with UK Standard 18 BACS payment validation. Both workflows run locally and preserve clean audit trails from trial balance to cash disbursement.
Troubleshooting context errors, size caps, and audit scope
Validation alerts list up to 20 diagnostic findings in the UI and export the full issue list into the Excel Validation sheet. Fixing these structural inconsistencies before submission avoids automated gateway rejections at national business registries.
Handling missing contextRef, unitRef, and duplicate warnings
Four diagnostic codes catch common tagging mistakes: MISSING_CONTEXT when a fact omits or references an unknown context ID, MISSING_UNIT when a numeric tag lacks a unitRef, NUMERIC_FORMAT when text cannot be parsed as a number, and DUPLICATE_INCONSISTENT when identical concept|contextRef|unitRef keys disagree by more than 0.001.
Guidance from the GOV.UK company annual accounts filing overview explains statutory filing paths for UK limited companies preparing tagged accounts. Spotting duplicate fact conflicts and unmapped unit references early helps audit teams trace mismatched rounding in disclosure tables before submission.
ZIP bomb limits, local worker privacy, and review scope
To protect browser tabs from compressed archive exhaustion, uploaded files are limited to 40 MB and extracted ZIP entries stop at 80 MB uncompressed. Processing runs 100% locally in a dedicated Web Worker with zero server transfers, keeping confidential earnings figures private before you verify Australian ABA direct-entry totals before bank submission.
Tip: Click Load sample iXBRL report to open
harbor-esef-2025-sample.xhtmlfirst. It demonstratesifrs-fulltags,scale="3", negativesign="-"finance costs, andSegmentsAxisdimensions so you can inspect the Excel export layout locally.
Extracted tables and validation messages are provided for informational purposes only and do not constitute legal, tax, accounting, or financial advice. This browser utility parses instance facts locally without fetching remote XBRL taxonomy linkbases or XSD schemas, so statutory filings still require formal taxonomy validation and review by a qualified accountant, chartered accountant, or tax advisor.