When we set out to build NetNerve, we recognized a major problem with existing cloud-based packet analysis platforms: trust. Packet captures (PCAPs) can contain incredibly sensitive information, from proprietary API keys to internal network topologies and even plaintext user credentials.
We knew that if security professionals were going to use NetNerve, we had to offer an airtight guarantee about data privacy. The solution was simple in concept but challenging in execution: never write the PCAP to disk.
The In-Memory Pipeline
Most file upload services work by receiving a multipart form request, saving the payload to a temporary directory like /tmp, processing it, and (hopefully) deleting it later. This is risky. If a process crashes before cleanup, the file remains. If the server is compromised, those temporary files are exposed.
NetNerve bypasses the filesystem entirely. When you upload a capture, it is streamed directly into memory buffer arrays. Our analysis engines — both the AI parser and the Suricata IDS — read directly from these ephemeral memory blocks.
No Database Retention
It's not just the raw files we protect. The metadata, the extracted payloads, and the AI-generated reports exist only for the duration of your session. We do not store your results in a PostgreSQL or MongoDB database. Once you close the browser tab or the session times out, the analysis is gone forever.
Server-Side Secrets Only
Additionally, we ensure that zero secrets are exposed to the client. The frontend is strictly a visualization layer. All communication with OpenAI, Stripe, and our Suricata container happens exclusively backend-to-backend. By keeping the frontend "dumb," we eliminate entire classes of client-side vulnerabilities.
The Bottom Line
Building a zero-retention platform required us to optimize memory usage heavily and build robust streaming architectures. But the tradeoff is worth it.
You can upload sensitive captures to NetNerve knowing that the moment you have your answers, the data is digitally incinerated.