Classification workload
Before
Four text-and-extension inputs repeated 10,000 times
After
40,000 runtime classifications completed in 1.2745 seconds
Direct answer
Foldora's September 6, 2026 runtime benchmark completed 40,000 schema-classification calls in 1.2745 seconds and discovered 1,001 eligible files in 0.0091 seconds. It ran on Windows 11 Pro with an Intel Core i5-8350U and 8 GB of RAM. These are runtime measurements, not a claim of perfect organization accuracy.
Reviewed and updated 2026-09-06
Results recorded on the test machine
The benchmark was executed from the current Foldora workspace immediately before these results were published.
- Runtime classifications
- 40,000 in 1.2745 seconds
- Eligible files discovered
- 1,001 in 0.0091 seconds
- Excluded dependency files
- 300
- Environment
- Windows 11 Pro, Core i5-8350U, 8 GB RAM
One machine and one synthetic fixture cannot represent every user's files. The raw timings are reported without converting them into a guaranteed speed or accuracy claim.
What this benchmark is designed to verify
The benchmark checks two bounded parts of the local workflow: repeated schema classification and recursive file discovery. Its discovery fixture creates 1,200 normal files plus 300 dependency files under node_modules, applies a 1,001-file limit, and verifies that the dependency directory is ignored.
It deliberately does not turn a synthetic timing result into a broad accuracy claim. Real folders contain larger documents, unusual formats, slower drives, and ambiguous context, so Foldora keeps the proposed result reviewable before changes are applied.
Step-by-step workflow
- Create four representative classification inputs covering an invoice, source code, a vacation photo, and meeting notes.
- Run those four inputs 10,000 times through the local schema classifier for 40,000 total classifications.
- Create 1,200 eligible files and 300 files inside an ignored node_modules directory.
- Run recursive discovery with a 1,001-file safety limit and record elapsed time and the returned file count.
Use-case examples
Discovery workload
Before
1,200 normal files plus 300 dependency files
After
1,001 eligible files returned in 0.0091 seconds; dependency files ignored
Evidence, meaning, and limits
Each measurement answers a narrow question so the result is useful without overstating it.
| Option | Best for | Tradeoff |
|---|---|---|
| Runtime classification | Checking repeated local schema decisions | Does not include full document extraction or model loading |
| File discovery | Checking bounded recursive scanning and ignored directories | Uses small synthetic files on one SSD-backed computer |
| Real-folder acceptance test | Evaluating whether suggested folders fit your own workflow | Requires human review because useful organization is contextual |
Frequently asked questions
Is 1.2745 seconds the time required to organize a folder?
No. It is the time for 40,000 runtime schema-classification calls. End-to-end time also depends on model startup, file reading, hardware, and the selected folder.
Does the benchmark prove 100% classification accuracy?
No. It measures repeatable runtime work. Foldora exposes a preview because ambiguous files still need a person to approve the proposed result.
Why publish the hardware?
Timings without an environment are difficult to interpret. The test used Windows 11 Pro, a Core i5-8350U, and 8 GB of RAM.
Related guides
Organize a folder with Foldora
Run local AI on Windows, review the proposed structure and filenames, then apply the changes you approve.
Download for Windows Get the complete bundle