PcapAI
pcapai.com
- Free plan
- Open source
- Windows
- macOS
- Linux
- Web
- API
- Self-hosted
What PcapAI is
PcapAI analyzes PCAP and PCAPNG network captures and creates security reports. Machine-learning models examine packet context and map findings to MITRE ATT&CK techniques. Downloadable PDFs include risk scores, findings, packet evidence, and remediation steps. The platform detects plaintext credentials, NTLM hashes, Kerberos tickets, and session tokens, then masks discovered credentials in reports. It also reports retransmissions, per-flow latency, TCP Zero Window events, and ARP conflicts. PcapAI accepts uploads without an account on its free homepage; compressed archives are unsupported and file-size limits vary by plan. Paid choices include a $9 one-time Single Report, monthly or annual subscriptions, and one-time credit packs. Advanced adds REST API access. An open-source MCP Server supports local analysis with AI assistants, while the maker says packet data stays on-device.
Who it suits
It is aimed at SecOps and SOC analysts, DevOps and NetOps engineers, and security consultants or pentesters. Its web, API, self-hosted, Windows, macOS, and Linux options support varied environments.
Good to know, and what to watch
- Maps findings to MITRE ATT&CK techniques
- Produces PDF reports with remediation steps
- Masks discovered credentials in reports
- Offers free uploads without an account
- Provides REST API on Advanced
- Compressed archives are unsupported
- File-size limits depend on plan
- No live capture is listed
- Paid plans limit report quantities
Plans
Every plan and priceSingle Report
Starter
Starter pack
Pro
Free plan
The verdict
PcapAI combines packet analysis, security findings, and troubleshooting metrics in downloadable reports. Its free upload option, local MCP server, and paid API tier give teams different analysis paths.
What the maker publishes
- Analysis speed
- The about page says a full report is typically produced in under two minutes. source
- API authentication
- API requests require a Bearer token and API access is restricted to Advanced subscribers. source
- API modes
- Uploads can specify security, troubleshooting, compliance, forensic or shadow-it analysis modes. source
- Audience
- The maker identifies SecOps and SOC analysts, DevOps and NetOps engineers, and security consultants and pentesters as its primary audiences. source
- Credential exposure
- It searches captures for cleartext passwords, NTLM hashes, Kerberos tickets and session tokens, then masks discovered secrets in reports. source
- Credential masking
- Cleartext credentials and sensitive payloads are mathematically masked in generated reports rather than printed. source
- Credentials
- The product detects plaintext credentials, NTLM hashes, Kerberos tickets, and session tokens, and masks discovered credentials in reports. source
- Developer documentation
- Official developer documentation covers MCP setup, REST authentication, upload polling, report download and automation examples. source
- File limits
- PcapAI accepts PCAP and PCAPNG files; compressed archives are unsupported, and file-size limits depend on the plan. source
- Free use
- The homepage invites users to upload any PCAP for free, and says no account is required. source
- Free use
- Users can upload a PCAP without an account and receive a free threat summary; the full forensic PDF costs $9. source
- GitHub macOS limitation
- The repository warns that its unsigned binary triggers macOS Gatekeeper and may require removing the quarantine attribute. source
- GitHub prerequisites
- The repository lists Rust 1.75 or newer, an optional PcapAI API key and Claude Desktop or another MCP-compatible client as prerequisites. source
- GitHub protocol
- The MCP server communicates using JSON-RPC 2.0 over standard input/output and uses reqwest with rustls plus Tokio. source
- GitHub repository status
- The public pcap-ai/mcp-server repository shows one star, zero forks and five commits when crawled. source
- GitHub size error
- The repository documents HTTP 413 errors when an uploaded capture exceeds the service limit. source
- GitHub workflow
- The repository documents a flow that reads a local capture, uploads it to the REST API, polls completion and saves the generated PDF beside the capture. source
- Integrations
- The Advanced plan includes a REST API, and the maker describes sending findings to Splunk, Elastic SIEM, or Microsoft Sentinel. source
- Integrations
- The maker describes use with SIEM systems, CI/CD pipelines, ticketing workflows and custom backend scripts. source
- Limit: capture scope
- The homepage says Wi-Fi monitor-mode, USB and Bluetooth captures are unsupported, and compressed archives are not accepted. source
- Limit: file size
- The free analyzer accepts files up to 20 MB; Starter allows 25 MB and Pro or Advanced allow 50 MB. source
- Limit: larger deployments
- Custom file-size limits, higher volumes and on-premise deployment require contacting sales. source
- Local analysis
- The maker offers an open-source MCP Server for local PCAP analysis with AI assistants such as Claude Desktop; it says packet data stays on the device. source
- MCP cloud dependency
- Although the MCP binary runs locally, it uploads the capture to PcapAI over TLS for server-side analysis and therefore does not operate offline. source
Checked against the maker's pages on 14 Oct 2026.