Please wait while we prepare your content.
Please wait while we prepare your content.
How SageTweaks writes guides, benchmarks performance, verifies safety claims, sources reviews, and corrects mistakes. Plain rules, no PR talk.
Every guide on SageTweaks is written by a named author — not anonymous staff, not AI alone. The author is listed at the top of the post, links to a public profile, and is accountable for every claim.
Guides are structured around concrete, reproducible changes — registry keys, services, power-plan settings, GPU-driver toggles. We don't publish vague advice like "tweak Windows to be faster." If we can't tell you what to change and why, we won't publish it.
When a guide quotes an FPS gain, that gain is from a benchmark on a specific machine, in a specific game, with the rig and Windows version stated. Variance and 1% lows are reported alongside averages so readers can judge whether the result generalizes to their hardware.
Our reference rigs are documented on the benchmarks methodology page so any reader can compare them to their own machine. Every test run records hardware, Windows version, GPU driver version, in-game settings, and game build version.
Each tweak is benchmarked in isolation before being combined. This is the only way to attribute a measured FPS change to a specific change rather than to test variance.
We use CapFrameX or PresentMon for frame-time capture, NVIDIA Reflex Analyzer for end-to-end input lag where available, and LatencyMon for DPC latency analysis. Tools are listed alongside results so the reader can reproduce the measurement.
When a claim cannot be benchmarked — for example, anti-cheat compatibility, where running the test in production would risk a ban — the article says so explicitly. We do not invent numbers to fill a gap.
Every public installer is submitted to VirusTotal and the report is linked from the home page and the dedicated /security page. Heuristic flags are named individually, with the reason each engine fires (Tauri/Rust binary heuristics).
The published SHA-256 lets readers verify the exact file before running it. The hash and the VirusTotal URL are derived from the same source-of-truth module so they cannot drift apart on the site.
Code-signing status is stated honestly. While the installer is unsigned, every page that mentions safety states that explicitly — we never claim a signed installer until the EV certificate ships and the next release is signed.
For comparisons of third-party tools, the same install/scan procedure is used on each tool so safety claims about competitors are held to the same bar as our own claims.
Reviews shown on the /reviews page and in homepage testimonials are real customer feedback. We never write or commission fake reviews and we never edit a review's content beyond formatting and minor grammar fixes.
Aggregate rating numbers are computed directly from the published reviews list — never inflated for SEO. If the review count is 27, the schema says 27, not 270.
Where a reviewer is anonymous by request, the review is still counted in the public list with the originally submitted text and rating; only the display name is generalised.
Optimization guides are reviewed after every major Windows feature update (24H2, 25H2) and after any major NVIDIA / AMD driver release that changes default behaviour. If a guide still applies, the modified date is bumped. If a guide no longer applies, it is rewritten or unpublished.
Game-specific guides are reviewed when the underlying game has a major patch that materially changes its performance profile.
Comparison content is re-tested at least once per year, and any time a competing product changes its pricing, packaging, or feature set in a way that materially affects the comparison.
If a reader spots an error — wrong claim, broken benchmark, outdated screenshot, missing caveat — email support@sagetweaks.com or open an issue at the repository linked from the support page. We aim to respond within 48 hours and to publish the correction within five business days.
Material corrections are noted at the bottom of the article with the original wording, the corrected wording, and the date. Typo and clarity edits are made silently.
If we are wrong, we say so. We do not memory-hole an incorrect claim.
SageTweaks is the publisher of this site and our own product. Comparisons of competing tools are written by reviewers who use a controlled hardware baseline; we are not paid by any of the tools we compare.
No SageTweaks page on this site contains paid placement or affiliate links at this time. If that ever changes, the article will carry a clearly labelled disclosure block above the fold.
We have not been paid for any external review or coverage we link to. Where another publication covers SageTweaks, that link is included only if the review was published independently of SageTweaks.
For corrections, factual challenges, or methodology questions, email support@sagetweaks.com with the URL of the article and the specific claim. Include a benchmark or a reproducible example where possible — that's the fastest path to a correction.
For security issues with the SageTweaks installer or website, see /security for the responsible-disclosure flow.