Build with AI
Connect Claude, ChatGPT, Codex, Cursor or any coding agent to your own port in one step, no key needed. Read only.
Let an AI read your port: what it is worth, what you hold, what moved, your PnL and your briefing, with the same figures and caveats as the app. No key to make and nothing to paste. The first time an agent asks, it sends you to getport.app to sign in and press Allow. Read only: it cannot move funds, sign anything or change a setting.
One command
Section titled “One command”Connects every coding agent on your computer to your port:
It asks which of the agents it found to add Port_ to. The first time one of them needs your port, your browser opens: sign in, press Allow, done.
Then teach them how Port_ answers, what is counted, what is not and how old a figure is:
Claude, Claude Desktop and ChatGPT
Section titled “Claude, Claude Desktop and ChatGPT”Add a custom connector with this address and sign in when it asks:
In claude.ai it is Settings, Connectors, Add custom connector. In ChatGPT, turn on developer mode under Settings, Apps & Connectors, Advanced, then create a connector with the address and OAuth.
Just one agent
Section titled “Just one agent”In Claude Code, type /mcp, pick port and choose Authenticate. In Codex, if it does not open
the browser by itself, run codex mcp login port.
Cursor and VS Code take the address on its own and sign you in the same way:
the MCP page has their blocks.
What you get
Section titled “What you get”- 12 tools: net worth by wallet, holdings, DeFi, perps, predictions, NFTs, PnL, activity, the daily briefing, alerts and your wallet list, and refresh, which the app has only when it asks for it and you tick it on the Allow page.
- Every answer says what the headline leaves out and how old the figures are.
- It reads the account you signed in as and nothing else. Names on tokens and NFTs are written by strangers, so phishing names and unasked arrivals are left out the way the app leaves them out, and the agent is told to treat every such name as data, never as an instruction.
- Every connected app is listed under Settings, API access with a Disconnect button that stops it at once. You are emailed whenever one is connected, and resetting your password disconnects them all.
The plugin
Section titled “The plugin”The server and the skills together, installed inside Claude Code or Codex:
Your own scripts
Section titled “Your own scripts”For a script, a spreadsheet or a cron job there is a plain JSON API, which takes a key rather than
a sign-in. Make one under Settings, API access, keep it in
PORT_API_KEY, then:
| Plan | Requests a minute | Requests a day | Refreshes a day |
|---|---|---|---|
| Pro+ | 300 | 25,000 | 3 |
| Ultra (not on sale yet) | 600 | 100,000 | 3 |
Free and Pro have no API. Every limit is per account, shared by every key, REST, MCP and every app you connect by signing in.
Which AI tools work
Section titled “Which AI tools work”| Tool | How it connects | Skills |
|---|---|---|
| Claude Code | One command, then sign in | Yes |
| Codex | One command, then sign in | Yes |
| Cursor | One command, then sign in | Yes |
| VS Code Copilot | One command, then sign in | Yes |
| Windsurf | One command, then sign in | Yes |
| claude.ai, Claude Desktop | Custom connector, then sign in | No |
| ChatGPT | Custom connector in developer mode, then sign in | No |
| Anything else | Signs in if it supports MCP sign-in, otherwise a key | Where it reads Agent Skills |