IM.codes CLI
A CLI links a machine to IM.codes for remote coding-agent sessions, shared memory and communication between agents
- 959 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- A bound machine exposes a remote terminal and project-file access
- Agents can run commands and send messages to other sessions
- The server holds history and binding keys; the shared evaluation instance is unsuitable for work secrets
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/im-codesDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- A bound machine exposes a remote terminal and project-file access
- Agents can run commands and send messages to other sessions
- The server holds history and binding keys; the shared evaluation instance is unsuitable for work secrets
This is third-party code. Review the repository files before installing.
What it does
IM.codes connects a work machine to a web dashboard for coding agents. The CLI configures a local daemon connecting the machine to the IM.codes server. A browser can continue Claude Code or Codex sessions, open a terminal and inspect file changes. For supported SDK connections, the daemon adds managed MCP tools. These provide shared memory and agent messaging, along with reminders and scheduled follow-ups to other sessions. The project owner selects collaboration and review participants. Shared Context reports MCP readiness for each supported connection. Check it before a task. Terminal sessions and SDK connections have different capabilities; support in one mode does not establish support in the other.
Who it is for. For developers continuing agent sessions from another device and sharing context between executors.
Good fit when
- You need to return to an agent session from a phone or another computer
- Several agents need shared project memory
- A neighboring session needs a message scoped to the current project
- Task execution needs a separate review of its result
Not a fit when
- You only need standalone memory MCP without a server and daemon
- You cannot allow remote terminal access to the work machine
- You require a managed service with commercial support and an uptime commitment
Example request
Prepare IM.codes on a separate machine for one Codex session. Show daemon access and available MCP tools before connecting a work repository.Limitations
Node.js 22 or later and a configured coding agent with its own model access are required. macOS and Linux use tmux, while Windows via ConPTY is experimental. Self-hosting needs Docker and a domain with DNS; setup starts PostgreSQL, the server and Caddy, creates an administrator and binds the local daemon. The shared app.im.codes instance is for evaluation and may rate-limit requests; do not send work secrets there.
How to disable. Stop tasks and the IM.codes daemon on the bound machine. Remove the machine binding and access in your dashboard, then run `npm uninstall -g imcodes`. Stop a self-hosted server separately and preserve needed history. Removing the dashboard does not undo project changes agents already made.
FAQ
How do I install IM.codes?
Run `npm install -g imcodes`, then use your server or the shared evaluation instance to bind the machine.
What does binding do?
`imcodes bind` connects a machine to the selected server, starts the daemon and registers it as a system service. The binding URL contains an access key and must not be published.
Must I manually add MCP to every session?
Not for supported SDK connections. The IM.codes daemon adds MCP and supplies current session and project context. Check readiness in Shared Context: a generic configuration snippet without that context is insufficient.
Does an auditing agent guarantee correct code?
No. Cross-agent review can surface issues, but assess the changes and executed tests. The executor environment determines process permissions.
Where do messages go when self-hosting?
Messages pass through your server to the work-machine daemon. Connected models and external tools may also receive data. If cloud memory sync is enabled, it adds another transfer path; hosting the dashboard yourself does not make the whole system local.
Does IM.codes include a model?
No. Model subscriptions and server resources are billed separately. IM.codes code uses MIT.
Discussion
No questions or reviews yet.
Related
CLI toolsA terminal workspace where several CLI agents like Codex, Claude and Gemini exchange messages and work on one task under human oversight
A tmux-based TUI that runs Claude Code, Codex and other agents in one list: status, quick prompts, worktrees and diff review
A local control panel for coding agents: it runs Claude Code and Codex, keeps a task queue and merges verified work into main.
A CLI connects installed coding agents into a team, dispatches work between sessions and collects results across machines.
A CLI that keeps coding agents' shared context in one place: instructions, skills, MCPs, subagents and durable memory across projects and machines
A local dashboard for running many agents across machines: memory, code graph and agent-to-agent messaging
README in short
IM.codes distinguishes terminal sessions from SDK connections. Beyond the terminal, its dashboard shows files and task schedules. Remote desktop is not available on every platform.