Hey ROS community,
We’ve been building something called **BlackBox** — think of it like a flight data recorder for your ROS 2 robots.
The idea: every time your robot does something unexpected in the field, you shouldn’t have to dig through raw logs hoping you can reproduce it. BlackBox lets you record **episodes** (a bounded window of what your robot was doing), tag failures, monitor `/diagnostics`, and see fleet-wide patterns over time — without major changes to your existing ROS 2 setup.
We’re genuinely at the stage where we want to know if we’re solving the right problem.
**How you can connect:**
- **ROS 2 node** — the primary integration, works today
- **REST API** — for non-ROS systems or custom pipelines
- **RabbitMQ & MQTT** — for event-driven and IoT setups *(under active development)*
If your stack doesn’t speak ROS 2 natively, the API/MQTT path is coming — and knowing what people actually need here would help us prioritize.
**Try it without signing up first:**
- Live sandbox with real demo data: BlackBox Robotics — Failure Intelligence for Robotics | ROS 2 Data Platform
- Analyze view: BlackBox Robotics — Failure Intelligence for Robotics | ROS 2 Data Platform
**Or create a free account and connect your own robot:**
Sign up at https://bbrobotics.in — it takes 2 minutes, no credit card. You can hook up a ROS 2 node, upload an episode, and see how it looks on your actual data. That’s probably the most useful way to poke holes in it.
**What we want from you:**
- Is this a problem you actually have in your deployments?
- What’s obviously missing that would make this useful?
- What would make you never use something like this?
- If you’ve dealt with hard-to-reproduce field failures, how are you handling it today?
- Would MQTT/RabbitMQ support change anything for your setup?
We’d rather hear “this is pointless because X” now than after 6 more months of building. Roasts, edge cases, and skeptical takes all welcome.