How to Control Claude Code from Telegram with Hermes Agent

Hermes Telegram Remote Control lets developers run Claude Code coding sessions from any Telegram chat using slash commands. /code starts a session, /status checks progress, /approve applies changes, and /stop terminates. Developers save 5-8 hours per week by completing coding tasks during commute or off-hours without opening a laptop.

A developer commuting 45 minutes each way loses 7.5 hours per week to travel time. Setting up SSH on mobile requires a VPN client, a terminal emulator, and patience with tiny keyboards. Most developers do not bother.

[ STAT ] 58% of developers report being unable to make progress on coding tasks when away from their primary workstation. Urgent fixes wait 4-8 hours and take twice as long due to context loss. — GitLab Developer Survey, 2024

The fix is not a better SSH client. It is eliminating the need for terminal access entirely. Telegram slash commands reduce the mobile coding interaction to three steps: open chat, type command, review result.

[TOOL: Hermes Agent] Middleware that bridges Telegram webhooks to local Claude Code processes. It receives commands, interprets intent, and manages session lifecycle. [TOOL: Claude Code] Code execution engine that receives prompts and produces diffs or debugging output. [TOOL: Telegram Bot API] Communication channel. Every command and response travels through Telegram's encrypted messaging layer.

The agentic decision happens in step three. When Hermes receives /code "add rate limiting to the API", it must decide whether this task is suitable for print-mode execution or requires an interactive session. It evaluates command specificity — single-command edits go to print-mode, multi-step instructions launch a tmux session.

For developers with 30+ minute commutes: you lose 5-10 hours weekly to transit time. Submit bug fixes and simple features during your commute and review diffs before you reach the office.

For on-call engineers handling production incidents: when an alert fires at 2 AM, diagnose and push a fix from your phone without booting a laptop. /code runs the fix, /status checks deployment, /rollback reverts if needed.

For freelance developers managing 3-5 client projects: you switch contexts 10-15 times daily. Each client gets its own Telegram thread with full session history. No more SSHing into the wrong server.

Bot Setup. Create a Telegram bot via BotFather and register the token with Hermes. Output: active bridge.
Channel Binding. Bind a Telegram chat to a project directory with /bind /path/to/project. Output: persistent mapping.
Task Dispatch. Send /code "add rate limiting to API" in the bound chat. Hermes parses the command and looks up the repo. This is the reasoning step: Hermes decides execution mode.
Execution. Hermes spawns Claude Code in the bound directory. stdout streams back to Telegram in real-time.
Diff Preview. Claude Code produces a diff. Hermes formats it as a message with add/remove line counts.
Approval. Reply to the diff with /approve or /reject. Hermes applies or discards changes.

20 minutes. That is the honest setup time if you have Node.js installed and a Telegram account.

Hermes Agent (latest) → Middleware bridging Telegram to Claude Code. Claude Code (CLI) → Code execution. Requires claude login. Telegram Bot API → Communication. Bot token from BotFather.

Gotcha: Claude Code's output uses ANSI escape codes for terminal formatting. Hermes must strip these before forwarding to Telegram, or your chat will display garbled characters. Check that your Hermes build has ANSI stripping enabled.

▸ Off-hours coding capacity: 0 hours productive during commute → 3-5 hours/week via Telegram ▸ Urgent fix turnaround: 4-8 hours from alert to fix → 15-45 minutes using /code + /approve from phone ▸ Context recovery time: 20-30 minutes rebuilding mental context → 2-3 minutes reviewing Telegram history ▸ First-week measurable: coding tasks completed entirely from mobile — target 5+ tasks in week one ▸ Infrastructure cost: $0 beyond existing Claude Code API quota

If the Hermes host machine loses internet, Telegram commands queue but do not execute until reconnection. No offline mode exists.
Complex multi-step debugging is impractical on Telegram's linear chat interface. Use this for targeted fixes only.
Anyone with access to the Telegram chat can see live output. Do not bind production-deploy chats to shared groups without restricting bot visibility.
(5 min) Create a Telegram bot: open Telegram, search for BotFather, send /newbot, and save the token.
(5 min) Install Hermes Agent: npm install -g hermes-agent and configure the Telegram bot token.
(5 min) Install Claude Code: npm install -g @anthropic-ai/claude-code and run claude login.
(5 min) Bind a test project: send /bind /path/to/test-project in the Telegram chat, then send /code "create a hello.py file".

Q: Is Hermes Telegram Remote Control secure? A: All commands travel through Telegram's encrypted channel. The Hermes bot token is the only credential needed to send commands. Keep the token private and restrict the bot to a private chat or a group with controlled membership.

Q: Can multiple developers use the same Telegram bot? A: Yes. Each developer binds their own chat-to-repo mapping. Hermes uses the chat ID to determine which project is being referenced.

Q: What types of coding tasks work best on Telegram? A: Single-command tasks work best: one-file edits, script creation, bug fixes with clear reproduction steps. Complex multi-file refactoring should be done from a terminal.

Q: What happens if I send /stop mid-execution? A: Hermes sends SIGTERM to the Claude Code process. Any uncommitted changes are discarded. The partial diff is saved for review.

Q: Do I need a static IP or open ports for this to work? A: No. Telegram uses webhook polling. Hermes connects outbound to Telegram's API. No open inbound ports or static IP required.