Chime vs Claude Code Remote Control
Anthropic shipped Claude Code Remote Control — driving a local Claude Code session from claude.ai or the mobile app. It's a real product with real docs, and it's genuinely useful for the audience it's built for. Here's how it compares to Chime, sourced from Anthropic's own documentation, not guesswork.
| Claude Code Remote Control | Chime | |
|---|---|---|
| Setup | claude.ai OAuth sign-in, on a Pro, Max, Team, or Enterprise plan | `chime pair` — keys generated locally, no account |
| Agent support | Claude Code only | Any agent with a shell or MCP — Claude Code, Codex, Cursor, aider, Gemini CLI, OpenClaw, Hermes, and more |
| Auth backend | claude.ai accounts only — API keys, Amazon Bedrock, Google Vertex, and LLM gateways are not supported for Remote Control sessions | Works however you already run your agent — API key, Bedrock, Vertex, a gateway, whatever's already wired up |
| Transport | Session traffic transits Anthropic's infrastructure; not end-to-end encrypted | End-to-end encrypted (Signal Protocol + PQXDH + Double Ratchet + ML-KEM-1024) — the relay only ever sees ciphertext |
| Self-host | Not available — Anthropic-operated only | One binary; run your own relay (`chime relay`) or use the managed one at chime.sh |
| Push / notify logic | "Claude decides" when to notify you — a model-driven judgment call | Deterministic — your agent explicitly pings you or asks for approval when your workflow calls for it |
| Multi-agent | One Claude Code session, mirrored to your phone | One thread per agent — run as many concurrently as you want, each with its own conversation |
| Cost | Requires an eligible Claude subscription (Pro, Max, Team, or Enterprise) | Free forever — chatting with your agents will never cost money, and self-hosting everything stays free |
Where Remote Control is the right choice
If you're already paying for Claude Pro or Max, run a single Claude Code session at a time, and want that exact session mirrored to your phone with no setup at all, Remote Control is the simpler pick — it's built into the product you're already using, and there's nothing to install. Chime is the better fit once you run more than one agent, run anything other than Claude Code, want the relay to be unable to read your messages, or want the option to self-host.
The gaps, in Anthropic's own words
Per Anthropic's Remote Control documentation: it requires signing in with a claude.ai account on an eligible plan — API keys, Amazon Bedrock, Google Vertex AI, and third-party gateways aren't supported for Remote Control sessions. Session traffic transits Anthropic's infrastructure to reach your phone, rather than staying end-to-end encrypted between your machine and your device. There's no self-hosted deployment option. And it works with Claude Code specifically — not Codex, Cursor, aider, or anything else you might be running.
None of that makes Remote Control a bad product — it makes it a different one, built to extend claude.ai, not to be a general-purpose, self-hostable, end-to-end encrypted messenger for every agent you run. That's the gap Chime fills.