Skip to content
Game save workbench
Local-first tools No installation Support status shown
DEVICE-LOCAL Favorites and recent pages stay in this browser.

Emulator & Retro Tools

N64 Controller Pak Forensic Compare

Compare two exact 32 KiB Nintendo 64 Controller Pak images read-only, map changed regions and raw index-word evidence, compare redundant metadata states and export a deterministic portable evidence bundle.

01 / Tool workbench

Use the runtime only when it is actually ready.

The application shell fails closed. An input interface is not enabled merely because a WordPress Tool record exists; the declared engine must be registered in the current build.

N64 Controller Pak Forensic Compare
BROWSER LOCAL

Checking whether the declared runtime module is loaded. No input has been processed.

Session change summary

No output has been built in this session.

This two-image workflow runs the exact Phase 49 N64 forensic report independently on Source A and Source B, then compares whole-image hashes, all four checksum-protected ID copies, both mirrored index pages, raw checksum-covered index-word evidence, bounded changed-byte ranges and the preserved Phase 48 recovery decisions. The ordered pair receives a deterministic pair ID. No directory or allocation-chain semantics are inferred and neither source is selected as preferred.

02 / Compatibility

Know the supported input before opening a file.

Compatibility fields are intentionally explicit. Missing values stay unclaimed rather than being filled with generic assumptions.

Accepted extensions

.mpk.bin

Platforms

Nintendo 64

Supported versions / editions

Standard 32 KiB Nintendo 64 Controller Pak / Mempak images already supported by the Phase 49 forensic report.

Input constraints

Both inputs must be exactly 32,768 bytes and individually satisfy the Phase 49 Controller Pak forensic input rule. Pair comparison remains read-only and does not turn disagreement into majority voting or a repair recommendation.

Known limitations

Read-only comparison only. Raw index-word values remain metadata evidence, not reconstructed directory/allocation chains. No majority voting, preferred-source selection, merge, chain/orphan repair, formatting, payload repair, simultaneous multi-region repair or repaired Pak output is authorized.

Output / result

Deterministic local comparison JSON, bounded changed-byte-range CSV, and a portable deterministic evidence-bundle JSON containing both Phase 49 reports, the comparison, CSV text, and a SHA-256 manifest for each embedded artifact. Neither Controller Pak source image nor repaired bytes are embedded.

03 / Instructions

Use a reversible workflow.

File-oriented workflows start with an untouched original. Tool-specific instructions add only details documented for this utility.

  1. 01Keep the original

    Preserve an untouched copy before creating or transforming binary output.

  2. 02Confirm scope

    Match the file size, format, platform, and version to the support information on this page.

  3. 03Process locally

    Use the runtime only when the page reports that its declared engine is loaded.

  4. 04Validate output

    Compare or test the result before replacing any working source file.

Tool-specific steps

Choose two original exact 32 KiB Controller Pak images. Preserve both sources. Export the comparison and evidence bundle to document what differs and how each source validates, without treating A or B as automatically newer, healthier or authoritative.

04 / Privacy and processing

The processing mode is part of the product contract.

SaveEditor.com should state local processing only when the implementation truly keeps file contents in the browser. Server or hybrid processing must be disclosed just as clearly.

Declared mode

Browser local

Runtime state

Operational

Tool-specific privacy notes

Phase 50 reads only the two explicitly selected local files in the browser. The forensic-compare-v1 runtime contains no fetch, XMLHttpRequest, WebSocket, sendBeacon, form upload, server-side save processing, or telemetry path. Comparison and evidence-bundle exports contain no raw source images, no repaired-save bytes, and never alter or replace either selected source file.

05 / Troubleshooting

When a file is rejected, do not force it.

A rejection can indicate an unsupported size, format, version, platform, or byte sequence. Preserve the original and correct the input instead of bypassing validation.

Wrong file or size

Re-check the documented input type and browser size limit before retrying.

Unsupported format

A generic converter does not imply support for every game-specific container, signing scheme, or serialization format.

Source was modified

Return to the untouched source instead of repeatedly processing a damaged or partially transformed file.

Known Tool issues

If valid redundant copies disagree or either source has multiple defects, keep the ambiguity visible. This Tool compares the evidence but deliberately does not vote, merge or reconstruct a directory or payload.

06 / Verification

Support claims should be traceable.

The verification date, method notes, and public references are kept separate from marketing language so compatibility can be reviewed over time.

Last verified2026-08-10
Profilen64-controller-pak-forensic-compare

Method notes

Verified with deterministic clean/equal/different/corrupt fixture pairs, independent SHA-256 and changed-byte-range calculations, exact parity with both Phase 49 source reports, stable ordered pair identifiers, deterministic JSON/CSV evidence exports, evidence-bundle artifact-manifest hash checks, immutable-source assertions, no preferred-source selection, no repaired-file output, source/installable parity, and fresh-extract release regression.

07 / FAQ

Questions to answer before editing.

These answers are generated from the Tool record and current runtime state rather than generic compatibility promises.

Is N64 Controller Pak Forensic Compare ready to use?
The Tool record is verified and declares an operational runtime. The page enables its input controls only when the declared engine module is loaded in the current build.
Does processing stay in the browser?
Yes for this declared mode. The loaded runtime processes the explicit input in the browser; Tool-specific privacy notes describe the exact scope and limits.
Should I keep the original file?
Yes. Keep an untouched source file until any generated output has been validated.
What if my input is rejected?
Use the validation message to correct the base, encoding, width, timezone, byte alignment, or size. The runtime does not guess through invalid input.

Related tools

Continue with the same tool family.

Only index-ready verified records from the same Tool Family are shown here.