SkillAgentSearch skills...

imessage-portal

Read-only MCP server for iMessage, reachable from anywhere over an embedded Tailscale tunnel with OAuth.

Install / Use

claude mcp add vinsidious -- npx -y github:vinsidious/imessage-portal

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

71/100

Supported Platforms

Claude Code
Claude Desktop
Cursor

Tags

Our assessment of imessage-portal

imessage-portal scores 71/100 on our quality scale, 873rd of 945 AI & Machine Learning skills we index.

Its MCP Server is 3.0 KB long, split into 5 sections with 3 code examples: a solid amount of guidance for an agent.

It has 3 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
26/30
Structure
16/20
Description
12/15
Adoption
3/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago, so imessage-portal is actively maintained.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 92/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-02. Automated pattern scan on 2026-10-02. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

imessage-portal compared with similar skills

All 4 of these similar skills score higher than imessage-portal; compare them before choosing.

SkillScoreStarsUpdatedFormat
imessage-portal (this skill)by vinsidious7133mo agoMCP Server
claude-memby thedotmack100118.0ktodayCLAUDE.md
Agent-Reachby Panniantong100112.8k10d agoCLAUDE.md
Understand-Anythingby Egonex-AI10093.0k2d agoCLAUDE.md
rufloby ruvnet10082.0ktodayCLAUDE.md

Frequently asked questions

How do I install imessage-portal?
Run claude mcp add vinsidious -- npx -y github:vinsidious/imessage-portal. The install tabs above show the steps for each supported agent.
Which AI agents does imessage-portal work with?
It is written for Claude Code, Claude Desktop and Cursor, as a MCP Server file. Other agents that read the same format can often use it too.
Is imessage-portal safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 92/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is imessage-portal still maintained?
The repository was last updated about 3 months ago, so imessage-portal is actively maintained.

iMessage Portal

A native macOS menu bar app that turns your iMessage history into an MCP server any AI agent can reach from anywhere — Claude Code, claude.ai, ChatGPT, Cursor — plus a plain-HTML web portal for humans and for agents that still don't speak MCP.

It is 100% read-only. The app keeps a live full-text index of the Messages database and publishes it at a stable public URL through an embedded Tailscale node, so your agent can search your texts, read a thread, and pull an attachment whether or not it runs on this Mac.

How it works

iMessage Portal.app (menu bar, SwiftUI)
 ├─ Indexer            chat.db → index.db (FTS5), incremental, every 3s
 ├─ HTTP server        127.0.0.1:4870
 │    ├─ /mcp          MCP server (Streamable HTTP + OAuth 2.1)
 │    └─ /messages…    plain HTML portal, token login
 └─ Embedded Tailscale (vendored libtailscale/TailscaleKit)
      └─ Funnel :443 → https://messages.<tailnet>.ts.net  (public, TLS)

No Tailscale app install, no BlueBubbles, no SIP changes. Onboarding is: grant Full Disk Access (macOS relaunches the app) → read the remote-access intro and click Continue → sign in to Tailscale in the browser (any SSO) and click Connect → approve Funnel on the consent page → you're online.

Connect your AI

The MCP address and access token are in the app's menu and settings.

Claude Code / anything that sends headers:

claude mcp add --transport http imessage https://messages.<tailnet>.ts.net/mcp \
  --header "Authorization: Bearer <access token>"

claude.ai, ChatGPT (custom connector / developer mode): add a connector with the MCP address. The client walks through OAuth: a consent page opens in your browser, and you approve with one click (or by pasting the access token). Tokens issued this way die whenever you regenerate the access token.

Tools: search_messages (FTS5 + person/date/chat filters), list_conversations, get_conversation, get_attachment, and the search/fetch pair ChatGPT's research modes expect.

Agents without MCP support get the same history as plain HTML: every page is a bookmarkable URL behind the same token (see /help on your portal). Ergo: MCP → M-See-P 🔥

Install

Grab the latest DMG from Releases, open it, and drag iMessage Portal into Applications. Builds are signed and notarized, so macOS opens them without any Gatekeeper warnings. The app updates itself from then on — it asks before checking automatically, and each update installs with one click. Apple Silicon only for now (the vendored Tailscale framework builds arm64).

Develop

make deps    # build vendored TailscaleKit.framework (Go + Xcode required)
make build   # swift build
make run     # run the menu bar app from the debug build
make test    # unit tests (PortalCore only)
make app     # assemble dist/iMessage Portal.app

Runtime data lives in ~/Library/Application Support/MessagesPortal/ (index.db, config.json with the access token, Tailscale node state, OAuth client/token store, attachment transcode cache).

Related Skills

View on GitHub
GitHub Stars3
CategoryAI
Updated2mo ago
Forks0

Languages

Swift

Trust signals

92/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 low