Should you trial Pangolin?
Pangolin combines reverse-proxy and VPN capabilities with granular access control. Consider a limited trial if you need both browser access to applications and installed-client access to private resources. Our failed source build leaves deployment readiness unresolved.
The README at our recorded commit describes these access paths, NAT traversal, and permissions assigned to specific resources. Those are documented capabilities. Our sandbox did not verify application access, authentication, client connections, or access-policy enforcement.
What our sandbox recorded
Our run on August 25, 2026, at commit 7319bf8 installed successfully but failed its source build. The full commit was 7319bf84f78c33d94490d24a85b5441d7db15d44. Installation returned exit code 0 in 111 seconds; the build returned exit code 1 after 30 seconds. Neither command timed out.
Our harness records command outcomes, exit codes, elapsed time, timeout status, and retained log tails. A successful installation command does not establish that a gateway started or served a workload. The build tail repeatedly references module-not-found diagnostics beside application page paths, but it does not name the missing modules or establish the cause. Peer-dependency warnings and pending install-script approvals also appear in the installation tail; their presence does not prove that they caused the build failure.
Tests were skipped because the harness detected no test script. The checkout contained a tests directory, so this is a limitation of the recorded execution rather than evidence that no tests existed. We have no passing or failing test count for this run.
No new execution was performed for this correction. Our sandbox did not deploy the published containers, connect sites or clients, exercise NAT traversal, benchmark throughput, or conduct a security review. The MrKeyoor testing methodology explains the reporting approach and its limits; its current description does not establish which methodology version the August run used. The attached sandbox image label and detected Node major also differ, so this record does not establish a verified Node runtime version for reproduction.
What the dependency record establishes
The attached stored parser summaries report 1,637 dependency packages occupying 1,580 MB and 14 npm audit advisories: 1 high, 3 moderate, 10 low, and none critical. These are the attached dated harness values, preserved without substituting the different totals in the raw installation footer.
The supplied evidence does not reconcile the footer with the summaries. Do not add their totals together or treat either package total as verified audit coverage. The audit record lacks scannedPackages, so scan coverage is unknown. The advisories are recorded dependency findings, not proof of exploitable flaws in a deployed Pangolin gateway. Stored parser summaries also do not establish unique or exhaustive test coverage.
Where public ports are still required
Self-hosting Pangolin requires a public gateway with open firewall ports. The quick-install guide specifies a Linux server with root access and a public IP, a domain pointing to it, an email address, TCP ports 80 and 443, and UDP ports 51820 and 21820.
The pinned README's description of outbound tunnels without public IPs or open ports concerns protected sites. It does not remove the gateway requirements. The installer pulls Pangolin, Gerbil, and Traefik images; Gerbil is optional for a standard reverse-proxy deployment without tunneling. These documented container steps were not part of our npm source run.
Which alternatives deserve comparison?
NetBird combines a peer-to-peer private network with centralized access control. Its current README also lists browser SSH and RDP, a reverse proxy with automatic TLS, and self-hosting. Compare it directly when those requirements overlap your Pangolin evaluation. The supplied evidence does not establish which platform is easier to operate or more reliable.
Tailscale provides private WireGuard networks. Its repository contains the daemon and CLI used by its clients. Consider it when private device connectivity is your main requirement. The supplied extract does not establish replacement coverage for Pangolin's browser publishing; a separately retained application proxy would complement that networking choice. We did not run either alternative.
What to verify before expanding a trial
The pinned README identifies Community Edition as AGPL-3 and Enterprise Edition as using the Fossorial Commercial License. It states that Enterprise Edition is free for personal and hobbyist use and businesses below $100K USD gross annual revenue. Check the applicable edition and license terms before deployment.
The update guide requires backups and explains that migrations can prevent downgrades; a failed database migration blocks startup. Start with a small resource set, verify permissions with separate user accounts, connect from the networks your users actually use, and test restoration before expanding access. These are recommended next steps, not completed sandbox checks.
How the scores are assessed
The editorial rubric runs from 1 to 5: 1 indicates a blocker without a demonstrated route through it; 2 indicates substantial unresolved friction; 3 indicates an operational path with material qualifications; 4 indicates strong evidence with limited gaps; 5 indicates comprehensive evidence within an explicitly bounded scope. Unknown dimensions receive no numerical score. These ratings are judgments about the supplied evidence, not benchmarks or adoption measurements.
Setup ease is 2 because our source build failed and the container installation remains untested here. Documentation is 5 within the supplied installation and update scope, which covers prerequisites, setup, migrations, backups, and verification. Community is unassessed because the extracts do not establish contributor participation or support outcomes. Maturity is 3 because operational procedures are documented, while our failed build and lack of deployment testing leave material qualifications.

