About PostAgent
PostAgent is an agent skill for sending documents to US postal addresses through a print-and-mail API. It supports several document formats and requires an explicit confirmation after providing a quote, before an irreversible paid send.
Key features
- 01Send PDF, HTML, Markdown, text, DOCX, and images
- 02Upload documents or mail-merge templates
- 03Quote address and postage details
- 04Pay per send in USDC on Base
- 05Track job status and USPS delivery
Use cases
- 01Mail a finished letter to a US address
- 02Send a mail-merge document
- 03Check a quote before authorizing postage
- 04Follow a mailed item's job and USPS tracking
MCP support
The skill works through the PostAgent REST API and can optionally use the PostAgent MCP server at `https://api.postagent.sh/mcp` when configured.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://api.postagent.sh/mcpclaude mcp add --transport http tech-interpretai-postagent https://api.postagent.sh/mcp{"mcpServers":{"tech-interpretai-postagent":{"type":"http","url":"https://api.postagent.sh/mcp"}}}Open in Cursor{"servers":{"tech-interpretai-postagent":{"type":"http","url":"https://api.postagent.sh/mcp"}}}Open in VS Code{"mcpServers":{"tech-interpretai-postagent":{"type":"http","url":"https://api.postagent.sh/mcp"}}}{"mcpServers":{"tech-interpretai-postagent":{"serverUrl":"https://api.postagent.sh/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
14| Parameter | Type | Required | Description |
|---|---|---|---|
| installedMcpVersion | string | No | Version advertised by the connected PostAgent MCP server, if known. |
| installedSkillVersion | string | No | Version from the installed PostAgent SKILL.md frontmatter, if known. |
Details
From the project READMEPostAgent — Print & Mail (Agent Skill)
An installable agent skill that lets your AI coding agent print and physically mail a document (PDF, HTML, Markdown, text, DOCX, or image) to a US postal address via the PostAgent API. Each send is paid per-call in USDC on Base using the x402 payment protocol, then handled by PostAgent's print-and-mail workflow.
Sending mail spends real money and is irreversible once submitted. The skill always quotes first and requires explicit user confirmation before paying.
Install
Claude Code (global) — always pass -a claude-code so it lands in
~/.claude/skills/, the only place Claude Code reads skills from:
npx skills add interpretai-tech/agent-tools --skill postagent-print-and-mail -g -a claude-code -y
Then restart Claude Code — the skill list is read once at session start, so a newly installed skill won't appear until you relaunch.
Other agents (auto-detected — installs to each agent's own skills dir, e.g.
.agents/skills/ for Codex/Cursor/Gemini):
npx skills add interpretai-tech/agent-tools --skill postagent-print-and-mail -y
Without
-a claude-code, the CLI may install only to the generic.agents/skills/path, which Claude Code does not read — the skill will appear "installed" but never load.
-yskips the interactive scope/confirmation prompt. Drop it if you prefer to confirm the scope and review before install.
What it does
- Upload a document (free) — finished letter or
{{field}}mail-merge template. - Quote (free) — verifies sender/recipient addresses and locks an inclusive USDC price for 15 minutes.
- Pay — the agent's x402 wallet pays the quote's
paymentUrl; PostAgent prints and mails the letter. - Track (free) — poll job status and USPS tracking.
Works in any agent via the REST API (plain HTTP, no setup) — and optionally
via the PostAgent MCP server (https://api.postagent.sh/mcp)
when it's configured. See SKILL.md for the full instructions the
agent follows.
Requirements
- A US sender and US recipient address (2-letter state code, 5-digit/ZIP+4).
- An x402-capable wallet with USDC on Base mainnet to pay per send.
License
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.