Multi-Run HAR Asset Budget Analyzer

Compare recorded HAR runs, review byte budgets and export a redacted summary.

On-device No sign-up Free to use

Compare HAR runs controls

Drop your files here

or choose them from your device

Two to five HAR files, at most 4 MiB each and 12 MiB total; supplied text is used when no files are selected.

    ?

    Up to five runs and 2,000 entries per run.

    Result

    Showing an example.

    Compare two sample app.js requests with the same 1,200 recorded body bytes and different durations, both below a 1,000,000-byte budget.

    2 runs compared; 0 exceed the entered known-byte budget.

    Loading the tool. Your result will appear here.

    https://usefreetools.net/multi-run-har-asset-budget-analyzer · 20 November 2026

    Was this helpful?

    How to use Multi-Run HAR Asset Budget Analyzer

    1. Choose two to five HAR files, or leave files unselected and supply a JSON array of runs. Selected files replace the pasted run array.
    2. Enter a whole-byte budget. If using the bytes/KiB helper, choose Apply transfer budget; editing its amount alone does not update the scenario.
    3. Review the updated run summaries, grouped request differences, missing sizes and recorded peak overlap. A total with unknown sizes is incomplete.
    4. Select Compare HAR runs to create the report/downloads. Inspect remaining hostnames, paths and MIME labels before sharing; the redacted report can still identify private resources.

    Example: Multi-Run HAR Asset Budget Analyzer

    Compare two sample app.js requests with the same 1,200 recorded body bytes and different durations, both below a 1,000,000-byte budget.

    You add
    HAR runs (JSON array): supplied example shown in the input; Entered transfer budget per run (bytes): 1000000
    You get
    2 runs compared; 0 exceed the entered known-byte budget.

    Options

    Transfer budget
    The budget is your limit. Known bytes can exceed it, while missing sizes prevent a complete total.
    Request matching
    Method plus normalized HTTP(S) URL identifies groups after query and fragment removal. Scheme, host, port and path remain; requests differing only by query can merge into one group.
    HAR run files (optional)
    Two to five HAR files, at most 4 MiB each and 12 MiB total; supplied text is used when no files are selected.
    HAR runs (JSON array)
    Up to five runs and 2,000 entries per run.

    Supported inputs and limits

    • Two to five runs, 2,000 entries each; files up to 4 MiB each and 12 MiB total.
    • Uses nonnegative response.bodySize observations, not decoded content.size or complete on-wire header/protocol overhead.
    • Missing/negative/invalid sizes remain unknown.
    • The entered budget is 0–1,000,000,000 whole bytes; helper KiB means 1,024 bytes.
    • Overlap counts recorded request intervals, not HTTP connections.
    More details

    No replay, Lighthouse/Core Web Vitals score or causal critical-path proof.

    Queries, headers and bodies are omitted from reports; remaining resource names may still be private.

    Where your input is processed

    This tool does not upload your inputs. See the limits for anything saved on your device.

    A redacted report can still identify private paths

    Headers, cookies, bodies, query values and initiators are omitted before export. Hostnames and paths remain useful for comparing assets and can still disclose internal names. Recorded overlap describes request intervals rather than HTTP connection count, and no network replay or performance score is produced.

    Questions about Multi-Run HAR Asset Budget Analyzer

    Are queries included in the report?

    They are removed from exported request identities.

    Does unknown size mean zero bytes?

    No. Missing transfer sizes remain explicitly unknown.

    Does concurrency count TCP connections?

    It counts overlapping recorded request intervals.

    Page updated · Project managed by Tony Hines · Report a problem with Multi-Run HAR Asset Budget Analyzer