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-portalIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
AI & Machine LearningSupported Platforms
Tags
Our assessment of imessage-portal
imessage-portal scores 71/100 on our quality scale, 837th of 932 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.
Maintenance, license and trust
- The repository was last updated 38 days 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.
imessage-portal compared with similar skills
All 4 of these similar skills score higher than imessage-portal; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| imessage-portal (this skill)by vinsidious | 71 | 3 | 38d ago | MCP Server |
| claude-memby thedotmack | 100 | 95.1k | today | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 87.5k | 16d ago | CLAUDE.md |
| Understand-Anythingby Egonex-AI | 100 | 84.9k | 3d ago | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 74.2k | today | CLAUDE.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?
- 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 38 days ago, so imessage-portal is actively maintained.
Skill content
View source on GitHubiMessage 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
claude-mem
95.1kPersistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Agent-Reach
87.5kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Understand-Anything
84.9kGraphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
headroom
74.2kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
