Results
Compare APIs
Put two or more APIs side by side — the same request, the same load — and see which is faster, which fails more, and whether their answers match.
Use Compare for questions like “is the new version faster than the old one?”, “is UAT as fast as production?”, or “which of these two services should we use?”
Set it up
- Open Compare in the sidebar. Each API gets a tab — you start with two, API 1 and API 2.
- Fill in each tab like a normal request — or Load saved request to load one into the open tab (the tab takes the request's name).
- Choose the load under Compare setup, below the request. It applies to every API, so their timelines line up.
Working with tabs:
- Add API adds a tab, up to 6.
- Double-click a tab to rename it; the name is used in the results and exports.
- Hover over a tab and click × to close it. A compare always keeps at least two tabs.
- After a Send, each tab shows the status code it got.
The Environment picker above the request fills {{variables}} the same way for every tab — see Environments. To compare UAT with production, give each tab its own address, for example https://uat.api.example.com/orders and https://api.example.com/orders.
Compare remembers your tabs and load between visits.
Past compares reopens any compare saved in History, with its full results.
Send all
Send all sends every API once and compares their answers: status, time, and the structure of the response — fields that are missing, extra, or of a different type. It is a quick way to spot a breaking change before you load anything.
Compare load
Compare load runs the performance test on every API. Run mode, under Compare setup, decides how:
- One after another — the fairest comparison: each API has your computer's CPU and network to itself;
- All at once — faster, but the APIs share your computer's capacity, so treat the numbers as a rough guide.
Before any load, every API is sent once. Each must answer with a 2xx, and by default all must return the same response structure. If they don't, the compare stops with Compare not started — responses differ. and lists the differences; Compare anyway goes ahead. To allow different structures every time — for example when comparing two different services — tick Allow different response structures under Compare setup. Each API still needs a 2xx.
While it runs, the results show each API's progress. When it has finished, Overall time taken shows the wall-clock time for the whole compare, including the checks and the gaps between runs. The compare is saved to History.
The results open full width, with a tab for each:
- Comparison — the key numbers side by side (the best value marked ★), the Result summary naming the winner on response time (p95) and on requests per second, each API's own result judged by your goals or the starting rules, and the charts overlaid so you can see where one pulls ahead;
- one tab per API with its full report.
Stop ends the compare early. When it has finished, Edit setup takes you back to the APIs; View results brings the results back.
Is it a fair comparison?
A run with less load or a shorter run has less to do, so it can look faster without being faster. When the runs had different load or length, the Comparison tab says so at the top. For example: Not a like-for-like comparison: API 3 ran 30 s with up to 5 users; API 1 and API 2 ran 2 min 43 s with up to 50 users. Then no winner is called and no value gets a ★. On the charts, a run that ended early is marked API 3 ended where its line stops.
Different requests are fine: comparing two APIs is the point.
Compare runs you already have
Collect finished runs like products in a shop:
- On any report, click Add to compare. A bar at the bottom of the page shows the runs you've collected (up to 6). Remove one with ×, or Clear them all.
- Keep browsing — the bar stays with you across pages, and even after a restart.
- Press Compare N runs. Compare opens the results with:
- Comparison — every run on the same charts, with the winner on response time and requests per second (when the runs are comparable);
- one tab per run with its full report.
Close returns to setting up a new compare.
The runs can be of different APIs and from different days — that's the point. When they test the same request, a run's own report also has a Before / after section.
In History you can also tick 2 to 6 performance tests and choose Overlay.
Save and share
Compares are saved in History; open one again with Past compares, or with Open overlay in History. The Comparison tab uses the Compare overview view, which you can customize like any report — see Views.
Export a compare as Excel, JSON, Response JSONs (each API's Send answer), or a Zip of every file, or share it — see Exports and sharing.