Restore / Research / Disavow File Format Disavow File Spec Disavow file format. The specification and the silent-failure cases.
UTF-8 / 7-bit ASCII .txt, max 100,000 lines, 2MB, 2,048 characters per URL. The domain: prefix disavows the entire referring domain. Comments via #. Format violations cause silent failure: the upload succeeds, the entries do not take effect.
The specification, one row per constraint
- Encoding: UTF-8 or 7-bit ASCII. Anything else causes silent failure.
- Extension: .txt. No HTML, no .docx, no exported .pdf.
- Maximum file size: 2 MB.
- Maximum line count: 100,000 lines.
- Maximum URL length: 2,048 characters.
- One URL per line.
- Comments: lines beginning with
#. - Domain-level disavow:
domain:example.comon its own line.
Worked example
A minimal valid file:
# Disavow file for example.com, generated 2026-05-28
# Format: Google Disavow Tool spec
# Source: Search Console + Ahrefs inbound profile audit
# Per-URL disavow entries (specific pages on otherwise-editorial domains)
https://low-quality-blog.example/post-1
https://low-quality-blog.example/post-2
# Domain-level disavow (entire referring site is the issue)
domain:pbn-network-1.example
domain:pbn-network-2.example
domain:scraper-aggregator.exampleOne file, all entries together. Submitted at search.google.com/search-console/disavow-links at the property level.
The silent-failure modes and the local validation step
The Disavow tool accepts the upload and reports success even when the file contains spec-violating entries. The violations do not take effect; the rest of the file does. The diagnostic surface for this is the reconsideration request itself, where Google's response can flag entries that did not apply. By that point the rejection has already cost a cycle.
Local validation before submission: encoding check (file reads cleanly as UTF-8 / ASCII), line-length check (no entries over 2,048 characters), no-HTML check (raw .txt only), domain: prefix sanity (no http(s) prefix on domain: lines), comment-syntax check (# only at line start). The audit gate before submission, not after rejection.
The engagement that runs this end-to-end for Restore clients is link disavow service: file authored to spec, validated locally, submitted with documented removal outreach in parallel for manual-action cases. The parent reference at Google Disavow tool covers the tool surface in full. Disavow meaning covers the operational role.
The diagnostic in front of the file step runs at the program.
What operators ask before submitting a disavow file.
- 01.
What is the maximum disavow file size?
2 MB total, 100,000 lines total, 2,048 characters per URL. Most operator-side files come in well under the line limit; the size ceilings are there for site-scale disavow runs.
- 02.
What does the domain: prefix do?
domain:example.comdisavows every URL on that registrable domain in a single line. Operationally faster than enumerating subdomain and path patterns; the recommended syntax when the entire referring site is the issue rather than specific pages. - 03.
What causes silent failure?
Encoding other than UTF-8 / 7-bit ASCII. Lines longer than 2,048 characters. Empty file. Non-text content (HTML wrapping, Word .docx, etc.). The Disavow tool accepts the upload but the spec-violating lines do not take effect. Validate locally before submission.
- 04.
Do I include the http(s) prefix on URLs?
Yes for per-URL disavow entries. Omit it on
domain:lines:domain:example.com, notdomain:https://example.com. The prefix on a domain: line silently invalidates the entry.
If your inbound profile reads as link-spam-shaped and you need the disavow authored to spec, book the diagnostic.
We pull the inbound profile, separate editorial / borderline / spam / orphaned, author the disavow file to Google's specification with the domain: prefix where the whole site is the issue, and run documented removal outreach in parallel for manual-action cases.