Software testing: a new line of work you can do from home
Companies leave their new builds here to be tested: websites, mobile apps, desktop software, embedded devices and games. You work through the steps from where you sit, on your own computer or phone. What you did is recorded as you go, and the value of the job is added to your balance.
| Type | Job | Cases | Slots | Fee |
|---|---|---|---|---|
| Card reader v2.3NewTakenPOS-4471Linux ARM, serial console 115200 | 24 | 4/5 slots | ₺1,850 | |
| Web | Basket and checkout flowMAG-2318Chrome, Safari, Firefox | 18 | 4/4 slots | ₺920 |
| Games | Chapter 3 save systemGME-3310PlayStation 5, real console | 26 | 3/8 slots | ₺1,980 |
| Mobile | iOS 17 sign-in screenAPP-7742iPhone 13 and 15, real device | 12 | 2/5 slots | ₺1,140 |
| Desktop | Windows installer wizardDSK-1180Windows 10 and 11, VM allowed | 9 | 1/3 slots | ₺680 |
- Slots open
- Slots full
- Yours
- 4job types: web, mobile, desktop, embedded
- Instanttaking a slot, with no approval step
- Weeklypayout frequency; the detail is on the payment page
- 24 mohow long a run record is kept
What we handle for you
Six things, all of them about one thing: that your work shows and gets paid on time.
The company opens the job, you don't hunt for it
A job enters the queue by itself whenever a company ships a new build. You write no proposal and negotiate no rate; the job is there, with its fee on it.
The queue is filtered to you
A job your level does not cover, that will not fit your free slots, or that needs an environment you do not have never appears at all.
Two separate slot counts
How many testers a job takes is on the job; how many jobs you can hold at once is set by your level. Both are known up front.
A run record you cannot edit
The environment fingerprint, the case timestamps and the file digests are written by the agent. Neither you nor we can change them afterwards.
Fixed fee, paid weekly
The amount is on the job and does not move after delivery. The value of an accepted job is added to your balance, and payouts run weekly.
Disputes run on the record
If a report is rejected, the argument runs on the record rather than on what either side remembers.
What you test
Five kinds of job arrive in the queue. Which ones you can take is decided by your level and by the device you actually have.
Web
Sites and admin panels. Browser matrices, form validation, checkout flows and accessibility cases. Your own machine is enough.
Mobile
iOS and Android apps. A real device is required, and the model and OS version the job wants are written on it.
Desktop
Windows, macOS and Linux software. Install, update and permission cases. A virtual machine is accepted.
Embedded
Card readers, meters, control boards. Physical device and serial console access are required, and these are the best paid jobs in the queue.
Games
PC, console and mobile games. Gameplay flow, saves, crashes and frame rate. Sessions run long, so the slot counts are wider.
One job, hour by hour
A firmware job landed in the queue at nine in the morning and entered the payment queue at noon. What happened in between:
BountyShift
Embedded: card reader v2.3
₺1,8505 slots, 3 taken
6 hours leftTake
The job lands
The manufacturer sent a new firmware build for the card reader. A 24-case job for card reader v2.3 entered the queue with five slots.
You take a slot
Your level covered embedded work and you had a slot free. Taking it was instant, and the job went to 4 of 5.
Environment verified
You connected the device and the agent recorded the firmware version, the serial console settings and the timestamp. That record cannot be edited afterwards.
You submitted your proof
The run record for all 24 cases went out in one package, with screenshots and console logs for the two bugs you found.
Accepted
The client approved the report. ₺1,850 was added to your balance.
You don't describe the work, you show it
A screenshot should not be all you have. The agent writes down what ran and when, and you add what you saw on top. If anyone disputes it, the record is on your side.
POS-4471 run record
- Environment verified Linux 6.6 ARM64, serial console 115200, firmware 2.3.0
- Case 7 of 24 Payment refused with no card inserted
- Evidence attached console.log (18 KB), screen-01.png
- Report submitted 24 cases run, 2 bugs reported
- Accepted Moved to the payment queue
What you earn
The fee is fixed per job and written on the job. Your level decides both how many jobs you can hold at once and which types are open to you.
| Level | Concurrent slots | Types unlocked | Acceptance rate needed |
|---|---|---|---|
| 1 | 1 | Web | — |
| 2 | 2 | Web, mobile | 85% |
| 3 | 4 | Desktop added | 90% |
| 4 | 6 | Embedded added | 94% |
What people ask
How are slots decided?
There are two separate slot counts. The job has its own: the client writes how many testers it takes, and it drops off the queue once full. And you have yours: your level sets how many jobs you can hold at the same time.
Does the number of bugs I find change the fee?
No. The fee is fixed per accepted job and written on the job. Finding zero bugs does not stop a report being accepted — what we pay for is that the cases were run.
How do I prove I ran the test?
The agent records the environment fingerprint and a start and end timestamp for every case by itself. You add screenshots, recordings, logs and reproduction steps on top. The two go in one package and cannot be edited afterwards.
What happens if my report is rejected?
The client has to give a reason. You can dispute it, and the dispute is judged on the run record. If it goes your way the fee is released and your acceptance rate is untouched.
Follow the queue from your phone
The notification arrives the moment a job lands. You take the slot and upload the proof from the phone — the screenshots and the video are already on it.
Apply, and the queue opens
We ask which types you have worked on and what devices you have. You start at level 1 and move up on your acceptance rate.