Match the report title.
Look for "RSC_007 referenced resource could not be found" or the closest title above. Keep the original EPUBCheck, KDP, or Kindle Previewer wording if it differs.
EPUB error guide
EPUB broken links and missing anchors
Scan when KDP, Previewer, or EPUBCheck says an EPUB link, image, stylesheet, chapter path, nav.xhtml entry, NCX target, or anchor is broken.
Best fit
A link target or anchor is missing.
Upload when the error mentions RSC_007, RSC_012, internal links with missing anchor, a missing image/CSS/chapter path, or a broken nav.xhtml or toc.ncx target.
Report shows
Source file, target file, and #id separately.
The scan names the file that contains the broken pointer, whether the target file exists, and whether the anchor after # exists inside that file.
Stop when
Choosing the destination would change reader meaning.
Do not auto-fix footnotes, repeated headings, absent images, missing chapters, moved files, or renamed anchors. Those require manual confirmation.
Scan when the message does not name the affected path. The report shows affected paths and repair decisions, not manuscript text.
No signup before the scan report.Checks EPUB package paths, not manuscript text.Repair download only after EPUBCheck passes again.
Need fixed-layout, DRM, or source-file help? Check unsupported or review-first EPUB cases.
Match these report titles
- RSC_007 referenced resource could not be found
- RSC_012 fragment identifier is not defined
- Internal link with missing anchor
- Internal links with missing anchor
- Broken internal link target
- Missing image, stylesheet, chapter, or anchor target
Use this guide when
Your message matches "RSC_007 referenced resource could not be found" or one of the report titles above.
Upload if
You have the exported reflowable .epub and need the affected path, file, field, or repair decision before editing.
Handle outside this tool
Use the source tool or publishing workflow for DOCX, PDF, KPF, KCB, fixed-layout, comics, image-first books, DRM, visual design, or KDP listing and approval problems.
Read the scan report
Decide if this report item matches.
Use the report label, affected file, and repair decision together. A matching title alone is not enough if the file path or EPUB structure points to a different problem.
A useful report does not stop at the error name. It connects the message to an internal file and a next action.
Check the affected file first.
Use the Affected area, Source file, Target file, or Problem type rows before editing OPF, nav, XHTML, CSS, or image paths.
Follow the repair decision.
Repairable means one clear structural change. Needs review or Not supported means use the named file, source project, or platform step.
Keep with your repair note.
Copy these fields from the report instead of rebuilding the fix from memory.
Source file
Target path or field
Original error
Repair decision
Next step
1. Example report output
What the scanner checks.
The report should show both sides of the link: the EPUB file that contains the reference and the target it tried to reach.
Source file
OEBPS/nav.xhtml
Target file
OEBPS/Text/chapter-03.xhtml
Anchor / fragment
The part after # in an internal link, such as #intro or #note-4.
Report may show
Target file exists: true. Target anchor exists: false.
Repair decision
Automatic only when one existing path or id differs by letter case, URL encoding, or both.
If the file is absent, moved, or renamed, restore or retarget it manually. Automatic repair is limited to a unique existing path or id with only case or URL-encoding differences.
2. Next step
Use the report to decide path repair, anchor repair, or manual review.
Broken-link errors look similar, but the next step changes depending on whether the file path or the anchor is missing.
Upload scan
You only have a vague KDP or Previewer message.
Use the scan to find the source XHTML, CSS, nav.xhtml, toc.ncx, or content.opf entry that contains the broken pointer.
Repair
One existing value differs only by case or URL encoding.
Align the href, src, CSS URL, nav link, NCX target, or anchor only when exactly one packaged path or id is equivalent after case or URL-encoding comparison.
Manual review
The target is absent or several matches look possible.
Stop when choosing the target would change a chapter link, footnote, image, note, or reading location.
3. Concrete path example
A broken link report separates the file from the anchor.
The file before # can exist while the anchor after # is still missing.
EPUBCheck: RSC-012: OEBPS/nav.xhtml links to Text/chapter2.xhtml#intro Report: Source file: OEBPS/nav.xhtml Target file: OEBPS/Text/chapter2.xhtml Target file exists: true Target anchor exists: false Fix decision: automatic only if one existing id differs by case or URL encoding; otherwise manual
Do not point the link to the top of the chapter just to clear the validator error.
Quick decision
Decide whether the broken link has one safe target.
A path mismatch can be a small structural repair. A missing note, heading, or chapter target can change where readers land.
Scan first
The message does not name the source file.
Upload the EPUB so the report can show where the broken reference lives and which target failed.
Safe fix
One path or id differs only by case or URL encoding.
Correct the reference only when exactly one existing packaged path or id is equivalent after case or URL-encoding comparison.
Stop
The reader destination is unclear.
Do not guess footnote targets, repeated headings, missing chapters, or absent images.
Start here
Start with the link target that KDP or EPUBCheck cannot open.
A broken EPUB link can be a missing file path or a missing anchor after #. Upload the EPUB when the message does not clearly show where the bad reference lives.
What it means
One file inside the EPUB points to another file or to an anchor that the package cannot resolve. OPF manifest means the file list inside content.opf; anchor or fragment means the part after # in an internal link.
Can it be fixed automatically?
A checked repair is possible when exactly one existing packaged path or target id differs from the reference only by letter case, URL encoding, or both.
What to check next
Find the file that contains the broken reference, such as a chapter XHTML file, stylesheet, nav.xhtml, toc.ncx, or content.opf.
What not to assume
It stops for manual review when the file is absent, several targets could match, the anchor belongs to a footnote or repeated heading, or the fix would require choosing where readers should land.
Common situations
Common broken-link cases.
Start from the kind of target that failed, then decide whether repair is safe.
A chapter points to a missing image or stylesheet.
The XHTML or CSS reference resolves to a file that is not packaged at that exact path.
Repair automatically only for one case or URL-encoding path difference. A moved, renamed, or absent asset needs manual restoration.
A TOC or footnote link points to a missing anchor.
The target XHTML exists, but the id after # is gone, renamed, or in another file.
Repair automatically only for one case or URL-encoding difference. A real rename or changed reading location is manual.
nav.xhtml or toc.ncx points to an old chapter path.
Navigation files often keep stale links after chapter renames or split/merge exports.
Update the navigation target only when the reading order still matches.
Broken-link messages to compare
Broken EPUB links: internal href, src, CSS url(), nav.xhtml, toc.ncx, or anchor targets cannot be resolved.
Where broken EPUB links usually appear
KDP, Kindle Previewer, or EPUBCheck reports missing resources, broken fragments, stale navigation links, missing images, missing stylesheets, missing chapters, or footnote anchors that do not exist.
What it means
The EPUB usually points to an old path, missing file, or missing anchor.
One file inside the EPUB points to another file or to an anchor that the package cannot resolve. OPF manifest means the file list inside content.opf; anchor or fragment means the part after # in an internal link.
Common causes include renamed chapter files, changed folder case, URL-encoded spaces, moved images, removed CSS files, nav.xhtml links left from an older export, toc.ncx targets that no longer exist, and heading or footnote ids regenerated by the source tool.
Before you edit
Check the source file, target file, and anchor separately.
Do not rename files or add anchors until the report shows whether the target file exists and whether the part after # exists inside that file.
- 1Find the file that contains the broken reference, such as a chapter XHTML file, stylesheet, nav.xhtml, toc.ncx, or content.opf.
- 2Split any link with # into the file path before # and the anchor or fragment after #.
- 3Check whether the target file exists at the exact EPUB path, including folder name, capitalization, spaces, and URL encoding.
- 4If the file exists, check whether the anchor id exists inside that target file.
- 5Treat footnotes, backlinks, repeated headings, and missing assets as manual review unless the report shows one clear target.
Why KDP checks it
Why KDP and EPUBCheck reject broken links.
EPUBCheck
EPUBCheck checks EPUB 2 and EPUB 3 files against the official rules and reports package, markup, link, and file-reference problems.
EPUB 3.3 package rules
The OPF file is the EPUB's map: metadata, file list, reading order, and navigation wiring. KDP and EPUBCheck use it to decide whether the book package is coherent.
Can this be fixed safely?
When a broken-link repair is safe.
When automatic repair is safe
A checked repair is possible when exactly one existing packaged path or target id differs from the reference only by letter case, URL encoding, or both.
When you need manual review
It stops for manual review when the file is absent, several targets could match, the anchor belongs to a footnote or repeated heading, or the fix would require choosing where readers should land.
Before / after example
Before: nav.xhtml links to Text/chapter-03.xhtml, but the package contains Text/chapter03.xhtml. After: the link can be aligned if that is the same chapter. Before: a footnote link points to #note-4 and no matching note is clear; that needs manual review.
Ready to retry?
Scan the EPUB for broken paths and anchors.
The scan checks chapter links, image references, stylesheet URLs, nav.xhtml links, toc.ncx entries, and anchors. It keeps other EPUBCheck issues visible in the same report.
Upload EPUB to scanFAQ
Questions authors ask about EPUB broken links.
Which broken links can EPUB Fixer repair?
Only a unique existing path or target id that differs from the reference by letter case, URL encoding, or both. Missing, moved, renamed, or ambiguous targets remain manual.
What is the difference between RSC_007 and RSC_012?
RSC_007 usually means the file path cannot be found. RSC_012 means the file exists, but the anchor or fragment after # is missing inside that file.
Does the scan hide other EPUB errors?
No. This entry focuses the report on link and anchor issues by default, but package, manifest, fatal, and unsupported issues still remain visible.
Should I add missing anchors by hand?
Only when you know exactly where the reader should land. Adding an anchor to the wrong heading or note can make the EPUB validate while sending readers to the wrong place.