WHOSEPORT
Pricing

Free during early access.

The CLI needs no account. The cloud dashboard is free while WhosePort is in early access.

Local
$0no account

The CLI, the local agent and the MCP server. Everything runs on your machine.

Install the CLI
  • Port, process, container and project discovery
  • Shared-port and health warnings
  • watch mode and --json output
  • whoseport ask --local with your own Anthropic, OpenAI or Gemini key
  • Read-only MCP server for coding agents
Cloud dashboardEarly access
$0during early access

Connect your machines and run investigations you can review and approve.

Create an account
  • Everything in Local
  • Organizations, projects and environments
  • Environment graph, ports, containers and recent changes
  • Investigations with evidence, hypotheses and diagnosis
  • Fixes you approve, executed by your agent and verified
  • Members with owner, admin and member roles
Included everywhere

The safety model is not a feature

Policy classes, approvals, verification and redaction work the same locally and in the cloud.

Read-only by default

Inspection never mutates. Policy classes every action before it runs.

Approval per mutation

Restarts and config edits wait for a yes. A changed .env file is backed up first.

Verified fixes

Port listening, HTTP 200, TCP reachable. Recorded with the investigation.

FAQ

Questions

Ask something else →
Nothing during early access. Pricing for paid plans is not set yet, and nothing is billed: there is no payment method to add.
Cloud investigations use the model provider WhosePort is configured with. Local investigations (whoseport ask --local) call Anthropic, OpenAI or Google Gemini with your own key, from your machine. Deterministic checks never call a model.
Nothing unless you connect the dashboard or ask a model. With the dashboard connected, only normalized, redacted evidence is uploaded: for .env files only variable names and host and port targets, never values; log lines with tokens stripped; no source code.
Not by default. At autonomy level 2 every change waits for your approval. Level 3, set on the machine, lets restarts of containers and known dev processes run without asking. There are no delete, prune or shell actions at any level.
The local agent always runs on your machine. The cloud side is not offered for self-hosting yet.

Ready to know what is running?

One command to install. Read-only until you say otherwise.

$bun add -g whoseport
Create an account →
Team dashboard

Environment graph, investigation history and approvals for everyone in your organization. Free during early access.

See pricing
Give it to your agent

One MCP config line for Claude Code, Cursor or any MCP client. Read-only tools, so it can look but never change anything.

$ claude mcp add whoseport -- whoseport mcp
MCP setup docs