The IRL streaming checklist to run before you go live
By Elan Harpaz · Published 2026-10-31
The ten minute check that prevents the four failures that end IRL streams: battery, data, bonding, and a relay you never tested.
Your encoder tops out at 1080p60. A higher-end encoder would go further.
Buy links are affiliate links: we earn a commission at no extra cost to you. Prices are tracked estimates, not live prices. Full disclosure
| Component | Pick | Price | |
|---|---|---|---|
| Camera | DJI Osmo Pocket 3 Gimbal cam with great low light, USB webcam mode, and DJI mic pairing. The IRL favorite. Buy on Amazon | $495 | Buy on Amazon |
| Encoder | Orange Pi 5 + belaBOX Dedicated HDMI-in encoder running the free, open-source belaBOX firmware with SRTLA bonding. The DIY backpack heart. Buy on Amazon | $150 | Buy on Amazon |
| Connection | Netgear Nighthawk M6 5G mobile router with an ethernet port, so you wire it straight to the encoder. Buy on Amazon | $447.22 | Buy on Amazon |
| Data Plan | Visible Unlimited Truly unlimited data on the Verizon network. The IRL favorite for a cheap single line. Learn more | ~$25/mo | Learn more |
| Bonding | belaBOX SRTLA Bonds several connections to your relay, built into the belaBOX encoder. Free. Learn more | Free (with belaBOX) | Learn more |
| Relay / VPS | belaBOX Cloud Managed SRT relay plus phone control to start/stop the encoder. The cheap, popular choice. Learn more | ~$10/mo | Learn more |
| Power Bank / Portable Battery | SHARGEEK 140 (20K, 140W) High-capacity, high-wattage bank that runs a power-hungry encoder all day. Buy on Amazon | $79.89 | Buy on Amazon |
| Audio | DJI Mic Mini Tiny wireless lav that pairs straight to the Osmo Pocket 3, keeping audio in sync. Buy on Amazon | $79 | Buy on Amazon |
| Mount / Stabilizer | DJI Osmo Mobile 6 Proven phone gimbal for smooth handheld footage with a built-in extension rod. Buy on Amazon | $99.99 | Buy on Amazon |
| Backpack / Case | Vivicreate Transparent Bubble Pack A transparent, heavily ventilated bubble pack (sold as a cat carrier) with standout airflow. Roomy and breathable for a hot-running rig if you embrace the look. Buy on Amazon | $21.99 | Buy on Amazon |
IRL streams rarely fail creatively. They fail because something in the bag ran out, fell over, or was never tested at the bitrate it was about to be asked for. This is the check worth running before you leave the house, in the order things actually go wrong.
Before you leave
- Every battery at full, including the phone you use for chat and the camera you assume is charged.
- Confirm the camera charges while outputting video, with the cable you are actually taking.
- Check the high-speed data left on the plan against the hours you intend to stream. The arithmetic is in our data usage guide.
- Start the stream at home and watch it on another device for a minute. A relay you have never tested is not a relay.
On the street, first five minutes
- Watch the bitrate graph, not the picture, for the first block you walk. Dips there tell you what the rest of the night looks like.
- If it is already dipping, drop a tier now. A stable 720p30 keeps an audience that a stuttering 1080p60 loses.
- Check audio on the stream itself, not in your monitor. Wind noise is inaudible under headphones and obvious to viewers.
What to carry that is not gear
- A short spare cable for every connection in the bag. Cables fail more than devices.
- Gaffer tape. Every rig develops a rattle or a dangling cable within an hour.
- A written note of your stream key or a way to reach it offline, for the night the app logs you out.
Frequently asked
What is the most common reason an IRL stream ends early?
Power, followed by data. Both are predictable before you leave, which is why they are the first two items on this list.
How do I test a relay before going live?
Start the stream at home on mobile data with your phone on wifi, and watch the channel on the phone. If the relay is working, briefly disabling the modem shows a holding screen rather than the channel going offline.
About this guide. Written by Elan Harpaz. We do not lab-test hardware. Every quality verdict on this page is a planning budget computed from a published model, and every price is a tracked estimate rather than a live one. How the numbers are derived: methodology. How picks are chosen: editorial policy.
Build on this in the configurator