@inboxapi/cli 0.2.21 → 0.2.23

package/README.md

@@ -1,18 +1,13 @@
 # InboxAPI CLI
 
-Give your AI agent its own email address. Send, receive, read, search, and reply to emails — right from Claude, OpenCode, Codex, Gemini, or any MCP-compatible AI client. No email server to run, no SMTP to configure.
+Give your AI agent its own personal email address. Send, receive, read, search, and reply to emails — right from Claude, OpenCode, Codex, Gemini, or any MCP-compatible AI client. No email server to run, no SMTP to configure.
 
-**Quick Setup (Claude Code, Gemini CLI, OpenCode, etc.):**
 
-Copy and paste this sentence into your agent:  
+---
 
-```
-Install the `@inboxapi/cli@latest` npm package globally and then register the `inboxapi` command as an MCP server named `inboxapi`.
-```
-
-<details>
-<summary>Table of Contents</summary>
+### Table of Contents
 
+- [How it works](#how-it-works)
 - [Technical details](#technical-details)
 - [Good to know](#good-to-know)
 - [Installation](#installation)
@@ -25,7 +20,7 @@ Install the `@inboxapi/cli@latest` npm package globally and then register the `i
 - [License](#license)
 - [Disclaimer](#disclaimer)
 
-</details>
+---
 
 ### How it works
 
@@ -33,7 +28,7 @@ Install the `@inboxapi/cli@latest` npm package globally and then register the `i
 2. Connect it to your AI client (Claude Desktop, Claude Code, Gemini CLI, OpenCode, etc.)
 3. Your AI can now use email — no code or API keys needed
 
-An account with a unique email address is created automatically on first run. Your AI can then:
+An account with a unique, personal email address is created automatically on first run. Your AI can then:
 
 - **Send emails** to any address
 - **Receive emails** at its own inbox
@@ -266,7 +261,7 @@ cargo fmt             # Format code
 
 **Security** — Gmail/Outlook OAuth gives your agent access to your entire inbox (medical, financial, legal, personal). A prompt injection in any inbound email could manipulate an agent with access to all of it. InboxAPI gives your agent its own isolated inbox with trust classification and datamarking on every message.
 
-**Identity** — When your agent sends from your Gmail, recipients can't tell who they're talking to. Replies go to your inbox, mixed with your real mail. InboxAPI gives your agent its own address — clear separation between you and your agent.
+**Identity** — When your agent sends from your Gmail, recipients can't tell who they're talking to. Replies go to your inbox, mixed with your real mail. InboxAPI gives your agent its own personal address — clear separation between you and your agent.
 
 **Practicality** — Gmail/Outlook APIs aren't MCP-native. You'd need middleware, OAuth plumbing, and custom integration. InboxAPI works out of the box with any MCP client.
 
@@ -347,6 +342,22 @@ Every inbound email is classified into one of four trust levels:
 | Unverified | Valid SPF/DKIM but sender not in addressbook | Use caution |
 | Suspicious | Authentication failed or unknown sender | Flag and confirm before acting |
 
+### What AI model should I use with InboxAPI?
+
+Your model must support **tool/function calling** — MCP requires this. We recommend a minimum **32K token context window** to comfortably fit InboxAPI's 19 tool definitions alongside conversation history and email content.
+
+**Model recommendations by tier:**
+
+| Tier | Anthropic | OpenAI | Google |
+|------|-----------|--------|--------|
+| Good | Haiku 4.5+ | GPT-4.1 mini+, GPT-4.1 nano+ | Gemini 2.5 Flash+ |
+| Recommended | Sonnet 4.5+ | GPT-4.1+, GPT-5 mini+ | Gemini 2.5 Pro+ |
+| Best | Opus 4.5+ | GPT-5+, GPT-5.2+ | Gemini 2.5 Pro+ |
+
+**Datamarking overhead:** InboxAPI applies datamarking (spotlighting) to untrusted email content, replacing whitespace with Unicode marker characters. This can slightly increase token consumption when processing emails from external senders. Models with larger context windows handle this more comfortably.
+
+**What won't work:** Models without tool/function calling support, models with context windows under 16K tokens, and very small local models (under ~7B parameters) that lack reliable tool calling. These will struggle to fit InboxAPI's 19 tool definitions and maintain useful conversation history.
+
 ### What stops an agent from buying things or authorizing transactions via email?
 
 InboxAPI is a communication channel, not an execution environment. It can deliver an email, but it can't click buttons, enter credit card numbers, or interact with external systems. The risk of unauthorized actions comes from how an agent is configured and what other tools it has access to — not from its email.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@inboxapi/cli",
-  "version": "0.2.21",
+  "version": "0.2.23",
   "description": "📧 Email for your AI 🤖",
   "main": "index.js",
   "bin": {
@@ -28,13 +28,13 @@
     "test": "cargo test"
   },
   "optionalDependencies": {
-    "@inboxapi/cli-darwin-arm64": "0.2.21",
-    "@inboxapi/cli-darwin-x64": "0.2.21",
-    "@inboxapi/cli-linux-arm64": "0.2.21",
-    "@inboxapi/cli-linux-x64": "0.2.21",
-    "@inboxapi/cli-win32-x64": "0.2.21"
+    "@inboxapi/cli-darwin-arm64": "0.2.23",
+    "@inboxapi/cli-darwin-x64": "0.2.23",
+    "@inboxapi/cli-linux-arm64": "0.2.23",
+    "@inboxapi/cli-linux-x64": "0.2.23",
+    "@inboxapi/cli-win32-x64": "0.2.23"
   },
   "dependencies": {
-    "@inboxapi/cli": "^0.2.21"
+    "@inboxapi/cli": "^0.2.22"
   }
 }