Saving a result
Every successful run is saved automatically. This is what gets stored, what the "You already ran this" prompt does, and why reopening never costs credits.
Last updated 2026-08-06
Summary#
There is no Save button, and you do not need one. When a tool finishes successfully, Metric Vault stores the result for you and keeps it for 90 days. You can reopen it any time from Saved Work at no cost, and if you try to run the same thing again the app offers you the saved copy before it spends anything. This page explains exactly what is stored, when it is stored, and what the reuse prompt is asking you.
Purpose#
Two things go wrong in research tools that do not save automatically. You lose work, because a browser tab closes and a paid report goes with it. And you pay twice, because a week later you cannot remember whether you already ran that domain, so you run it again.
Automatic saving removes both. Every run becomes a permanent, reopenable artefact the moment it succeeds, and the platform is the one that remembers what you have already paid for. That is also why reopening is free: you have already bought that answer, and charging you to look at it again would be charging you twice for the same work.
Requirements#
- A signed-in account. Anonymous runs on the public free tools are not saved.
- A run that actually succeeded. A run that errors is not saved as a result, and neither is an empty answer from a provider.
- Nothing else. Saving is not gated by plan and does not consume credits.
Permissions#
Every signed-in user has Saved Work, on every plan including Free. Saved results are private to the account that ran them. Nobody else in your team can open, edit or delete your saved results, and deleting one requires your own signed-in session rather than just your email address.
Navigation Path#
You do not go anywhere to save. To find what has been saved:
Dashboard → Library → Saved Work
Step-by-Step Guide#
- Run any tool as normal. See How to run a tool if you have not yet.
- Wait for the result to finish rendering. The result is saved as soon as the run succeeds, before you do anything with it.
- Let the page settle for a couple of seconds. A moment after the report has painted, the app captures the rendered report and attaches it to the saved row. This is what makes a reopened result look identical to the live one instead of being redrawn from raw data.
- Check it landed. Open
Library → Saved work. The newest run appears at the top of its tool group, tagged Free to reopen. - Reopen it whenever you need it. Press Open on the card. The report opens in a full-width viewer with the badge No credits used. See Saved Work.
What is stored in a saved result#
| Stored | Detail |
|---|---|
| The tool | Which tool produced it, shown as its friendly name |
| The target | The domain, URL, keyword or text you ran it on |
| The settings | The country, date range and other options in force at the time |
| The cost | What the run cost in credits when you ran it |
| The timestamp | The exact date and time of the run |
| The data | The full result data, which is what Download JSON gives you |
| The report | The rendered report exactly as it was drawn on screen |
Tools that save differently#
Most tools save on the server the moment the run succeeds. Three tools do their work entirely inside your browser and never upload your file, so they record the run themselves once it finishes: Image Resize, Image Reformat and Responsiveness. Image runs keep the produced picture with them, which is why their saved viewer offers Download image.
The "You already ran this" prompt#
Before spending credits, the app checks whether you have already run this exact thing. A match means the same tool, the same target and the same settings, saved within the last 90 days. When one is found you get a dialog instead of a run:
- Heading: You already ran this
- Sub-heading: Open the saved result for free, or run it again for fresh data.
- Three rows of context: Query, Run (
today,yesterdayorN days ago) and Cost then - Three choices: Open saved result, Run fresh (N credits) and Cancel
Pressing Escape or clicking outside the dialog cancels it, and nothing is charged.
Note: The number on Run fresh is the tool's cost today, not what you paid last time. If a tool's cost has changed since your last run, the button shows the current price.
The prompt is deliberately skipped in two situations: when a workflow is running several tools at once, because the workflow already confirmed its total up front; and when you press Get recommendations, because that request is built from the result rather than typed by you. See Get Recommendations.
Reopening versus running fresh#
| Reopen | Run fresh | |
|---|---|---|
| Cost | Free, always | The tool's current credit cost |
| Data | Exactly what you saw at the time | Newly fetched |
| Speed | Immediate | A normal run |
| Use it when | You want the same answer again, or a record of what it said | The numbers may have moved |
If you only want current numbers and do not care about the old ones, run fresh. If you are assembling a client deliverable, comparing two dates, or defending a figure you quoted last month, reopen the saved copy.
Retention#
Saved results are kept for 90 days from the run, then removed by a scheduled clean-up job. This is a hard limit, it is the same on every plan, and it is not extended by opening a result. Cancelling a subscription does not delete anything early; the 90-day clock is the only thing that removes a saved result apart from you deleting it. See Data retention and deletion.
If you need a result to outlive 90 days, export it. A downloaded PDF, Excel, CSV or JSON file is yours permanently. See Exporting results.
Size limits#
Two limits can affect what is kept, and both fail gracefully.
| Limit | What happens |
|---|---|
| A result payload larger than about 400 KB | The run works and renders normally, but is not written to Saved Work |
| A rendered report larger than about 900,000 characters | The run is saved and reopenable, but replays from the data rather than from the captured report |
Neither costs you anything extra, and neither produces an error on screen. If a very large result matters to you, export it straight away rather than relying on finding it later.
Examples#
Example: On Monday you run Domain Overview on nike.com for the United States. It costs 6 credits and saves. On Thursday a colleague asks for the same report, so you type nike.com again and press Run. The prompt appears: Run reads 3 days ago, Cost then reads 6 credits. You press Open saved result, export it as a PDF and send it. Total additional cost: nothing.
Example: Two weeks later the same client asks whether their traffic has moved. This time you press Run fresh (6 credits), because the point of the question is the change. Both runs are now in Saved Work, so you can open the old one beside the new one and quote the difference.
Screenshots#
Tips#
- Give a run a recognizable target. The saved card is labelled with what you typed, so acme.com/pricing is far easier to find later than a bare homepage.
- Reopen before you re-run. The prompt only appears when the settings match exactly; changing the country or the date range makes it a different run.
- Export anything you will need after 90 days, on the day you run it.
Best Practices#
- Treat Saved Work as your working archive and exports as your permanent one.
- When preparing a client deliverable across several tools, run everything first, then reopen and export in one pass. Reopening is free, so batching costs nothing.
- If you are comparing periods, keep both runs rather than overwriting your memory of the first.
Common Mistakes#
- Assuming a failed run was saved. Only successful runs are stored. If the result area showed a red error, there is nothing in Saved Work.
- Pressing Run fresh out of habit. The prompt exists to save you credits; the primary action is the free one.
- Relying on saved work for long-term records. Ninety days is the limit. Export what you must keep.
Limitations#
- Saved results are per account, not per workspace. A teammate cannot open yours.
- There is no way to rename, tag or annotate a saved result.
- There is no way to extend the 90-day retention.
- Runs made on the public free tools while signed out are not saved anywhere.
Troubleshooting#
| Symptom | Likely cause | Fix |
|---|---|---|
| A run does not appear in Saved work | The run errored, or the result exceeded the storage size limit | Re-run it and confirm it finishes cleanly; export large results directly |
| The prompt did not appear even though you ran this before | A setting differs, such as country or date range, or the earlier run is more than 90 days old | Check the topbar country and range, then decide whether you still need a fresh run |
| A reopened result looks plainer than the live one | The rendered report was not attached, so the viewer redrew it from the data | Use Run fresh if you need the full styled report, or export the live result next time |
Sign in to view your saved work | Your session ended | Sign in again and reopen the Library |
FAQs#
Do I have to save anything myself? No. Every successful run is saved automatically. There is no Save button anywhere in the app.
Does reopening a saved result cost credits? No. Reopening is always free, which is what the Free to reopen badge on each card means.
How long do saved results last? Ninety days from the run. After that they are deleted by a scheduled clean-up job.
Can a teammate open my saved results? No. Saved results are scoped to the account that ran them. See What teammates share for what is shared and what is not.
What is the difference between this and the shared cache? The shared cache is a platform-level optimization that can return recent provider data quickly; it does not change what you are charged. Saved Work is your own copy of a run you already paid for. See Result caching and freshness.
If a run fails after the credits are taken, is it saved? No. A failed run produces no saved result. Credits are metered before the provider is called, so a failure can still consume them. See How credits work and A tool failed to run.
See also
Was this article helpful?
Thanks — feedback noted for the docs team.