First Task: Detect and Record
In the quick start you confirmed AI detection works in the live preview. This tutorial turns that detection into something the device does on its own: when a target appears, reCamera Pro records it automatically — and you play the clip back from the Web UI.
The workflow:
- Prepare: pick a model and a target
- Set the recording format
- Configure the AI inference trigger
- Apply the rule and trigger a recording
- Find and play back the recording
1. Prepare: pick a model and a target
- Make sure AI inference is running: AI Inference → Inference Configuration → Inference Enable on, a detection model selected as the Running Model, status Running. If you have not done this yet, follow step 4 of the quick start.
- Decide what should trigger the recording. The simplest first task: record when a person walks into view. The available categories come from the running model's output (e.g. Person, Vehicle, Hardhat, Mask, Safety Vest, Machinery).
The detection parameters themselves (categories, confidence/IOU thresholds, inference frequency) are configured on the AI Inference configuration page — this tutorial only references them.
2. Set the recording format
Open Recording Settings in the left navigation bar. At the top of the page, select the recording file format:
| Parameter | Description |
|---|---|
| Recording Format | Format for saved files: MP4 (video), JPG (images), or RAW (original image data) |
For this tutorial, choose MP4. Note: when MP4 is selected, each recording has a fixed duration of 1 minute.
All recording options on this page — trigger methods, schedules, storage quotas — are documented in detail on the event recording configuration page.
3. Configure the AI inference trigger
In the trigger method list, choose AI Inference Trigger and click Configure:
| Configuration item | What to set for this tutorial |
|---|---|
| Inference Task | Select the running AI inference task |
| Confidence Range | Keep the default range for your first recording |
| Category Filter | Select Person (or the category you want to detect) |
| Trigger Area | Leave empty — with no region drawn, the whole frame triggers |
Optionally, use Schedule Management to limit when the rule is active (e.g. night hours only). For a first test, leave the schedule covering the current time.
4. Apply the rule and trigger a recording
- Click Apply next to AI Inference Trigger. The status at the top of the page shows the active trigger method, e.g. Current: AI Inference Trigger.
- Walk into the camera's view (or hold up your chosen target).
- The system detects the target and automatically starts recording. With MP4 format, the clip is 1 minute long.
5. Find and play back the recording
Still in Recording Settings:
- Open File Preview — recorded files are listed in directories by date.
- Open the file for today: the Web UI plays it inline, so you can review the event without downloading anything.
- Check Storage Management to see disk usage and where files are saved. You can set a storage quota and enable overwriting old files so the latest recordings are always kept.
Storage behavior in depth (quotas, full-disk policy, download): see Find recordings and storage.
Result
You now have a standalone AI recording rule: no computer involved, the device detects and records on its own. Typical next steps:
- Narrow the trigger to a zone (entrance, hazardous area) with Trigger Area.
- Switch or combine trigger sources — scheduled, GPIO, serial, HTTPS, loop, or sound events — on the event recording configuration page.
- Send detection results out to your own system via MQTT or HTTP/UART.
- Before mounting the device permanently, run the pre-deployment checklist.
Tech Support & Product Discussion
Thank you for choosing our products! We are here to provide you with different support to ensure that your experience with our products is as smooth as possible. We offer several communication channels to cater to different preferences and needs.