Connect your AI

Postware connects to AI assistants and AI workflows over MCP. The fastest way is to paste the prompt from Agent setup, and your agent does the rest. Use this page to connect a tool by hand.

Every workspace has its own MCP server. Replace YOUR-WORKSPACE with your workspace's address:

https://api.postware.ai/w/YOUR-WORKSPACE/mcp

The workspace's Connect page in the dashboard has this URL ready to copy, and creates keys. There are two ways to sign in:

Treat a key like a password. Keep it in the tool's own settings and never commit it.

AI assistants

Claude Code

claude mcp add --transport http postware "https://api.postware.ai/w/YOUR-WORKSPACE/mcp" --header "Authorization: Bearer YOUR-KEY"

Add --scope user after add to use it from any folder.

Claude

1. Go to Customize, then Connectors, then Add custom connector. On Team and Enterprise, an owner adds it under Organization settings, then Connectors.
2. Paste the URL. If it asks how to sign in, choose Register automatically.
3. Sign in with your email and press Allow.

ChatGPT

1. Go to chatgpt.com/plugins, press the plus button and choose Add custom MCP server.
2. Paste the URL and choose OAuth for authentication.
3. Create it, install it, then sign in with your email and press Allow.
4. Type @ in the prompt box to use it.

OpenAI decides which plans can add custom servers and whether they can write. Check OpenAI's guide for your plan.

Cursor

Add this to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "postware": {
      "url": "https://api.postware.ai/w/YOUR-WORKSPACE/mcp",
      "headers": { "Authorization": "Bearer YOUR-KEY" }
    }
  }
}

Codex

Set POSTWARE_API_KEY to your key in the shell that starts Codex, then run:

codex mcp add postware --url "https://api.postware.ai/w/YOUR-WORKSPACE/mcp" --bearer-token-env-var POSTWARE_API_KEY

VS Code

Run MCP: Open User Configuration from the command palette and add this. VS Code asks for the key once and stores it securely.

{
  "inputs": [
    { "type": "promptString", "id": "postware-key", "description": "Postware API key", "password": true }
  ],
  "servers": {
    "postware": {
      "type": "http",
      "url": "https://api.postware.ai/w/YOUR-WORKSPACE/mcp",
      "headers": { "Authorization": "Bearer ${input:postware-key}" }
    }
  }
}

Windsurf

Windsurf is now Devin Desktop. In Cascade, choose Open MCP config file, which opens ~/.config/devin/mcp_config.json, and add this:

{
  "mcpServers": {
    "postware": {
      "serverUrl": "https://api.postware.ai/w/YOUR-WORKSPACE/mcp",
      "headers": { "Authorization": "Bearer YOUR-KEY" }
    }
  }
}

Gemini CLI

gemini mcp add --transport http --header "Authorization: Bearer YOUR-KEY" postware "https://api.postware.ai/w/YOUR-WORKSPACE/mcp"

AI workflows

Give each workflow its own key, limited to what it needs. You create keys on the workspace's API keys page.

n8n

Add the MCP Client Tool node to your AI Agent and set:

Claude Agent SDK

const options = {
  mcpServers: {
    postware: {
      type: "http",
      url: "https://api.postware.ai/w/YOUR-WORKSPACE/mcp",
      headers: { Authorization: `Bearer ${process.env.POSTWARE_API_KEY}` },
    },
  },
  allowedTools: ["mcp__postware__*"],
};

OpenAI Agents SDK

from agents import Agent
from agents.mcp import MCPServerStreamableHttp

async with MCPServerStreamableHttp(
    name="postware",
    params={"url": url, "headers": {"Authorization": f"Bearer {key}"}},
) as server:
    agent = Agent(name="Assistant", mcp_servers=[server])

Vercel AI SDK

import { createMCPClient } from "@ai-sdk/mcp";

const client = await createMCPClient({
  transport: { type: "http", url, headers: { Authorization: `Bearer ${key}` } },
});
const tools = await client.tools();

LangChain

Use @langchain/mcp-adapters in JavaScript or langchain.mcp in Python, with the same URL and header.

Anything else

Any tool that can reach a remote MCP server over Streamable HTTP works. Give it the URL and the Authorization header.

Tools

Every workspace has these 11 tools. Start with post_schema_describe, so your agent learns the models before it writes.

Install the skill

MCP connects your agent. The skill teaches it how to use your data well: where each kind of data goes, how to avoid duplicates and when to test a change first.

For Claude Code, install it for every folder with:

mkdir -p ~/.claude/skills/postware && curl -fsS https://postware.ai/docs/skill.md -o ~/.claude/skills/postware/SKILL.md

For other agents, add the contents of https://postware.ai/docs/skill.md to AGENTS.md or the agent's global instructions. Chat apps do not need it, because the server tells them how to use the workspace when they connect.