double message telegram

[meager heart]

Help double messages on telegram from the bot! Revoked API key. Hard kill daemon restart everything bunch of times. Still getting it. Only polling no webhook. Strange.

[spare flint]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[deft pulsar]

Based on my knowledge of Clawdbot and Telegram polling issues, here are the most likely causes and fixes for double messages:

Common Causes & Solutions

1. Two Bot Instances Running (Most Common)

Even after hard kill, a process might still be hanging.

# Check for any clawdbot/telegram processes
ps aux | grep -i clawdbot
ps aux | grep -i telegram

# Check for processes on your telegram port
lsof -i :18789

# Kill any lingering processes

pkill -9 -f clawdbot
pkill -9 -f telegram

2. Polling Offset Issue

When using polling, the offset might not be advancing properly. Check your gateway logs:

clawdbot logs --follow | grep -i "telegram\|polling\|offset"

3. State/Cache Corruption

Try clearing the state files:

# Clear any Telegram-related state
rm -rf ~/.clawdbot/state/telegram*
rm -rf ~/.clawdbot/cache/telegram*

# Restart fresh
clawdbot daemon restart

4. Update to Latest Version

There was a recent update that improved Telegram message handling:

cd ~/clawdbot-ref && git pull
clawdbot update
clawdbot daemon restart

5. Check Configuration

Make sure you're not accidentally running both polling and webhook simultaneously (even if webhook is "disabled", there could be a config conflict):

clawdbot config get telegram

---

Try this sequence:

1. Full daemon stop: clawdbot daemon stop
2. pkill -9 -f clawdbot
3. Clear cache/state as above
4. clawdbot daemon start
5. Watch logs: clawdbot logs --follow

If it persists after all that, share the relevant log output when the duplicate occurs — the process message entries will show if it's being handled twice.

[sterile dragon]

altually repToMode: off

[deft pulsar]

With replyToMode: off, double messages are usually caused by one of these:

Quick Fixes to Try

1. Update to Latest (there were telegram refactors Jan 14):

cd ~/clawdbot-ref && git pull
clawdbot update
clawdbot daemon restart

2. Check your actual running version:

clawdbot --version

3. Look at the logs right when it happens:

clawdbot logs --follow | grep -E "process message|dispatch|chatId=${your_chat_id}"

The process message logs should show if the same message is being handled twice. If you see two process message dispatching entries for the same messageId, it's a handler issue. If you see two process message complete entries, something's triggering two dispatches.

---

What to share:

• Output of clawdbot --version
• The log lines around the duplicate (look for same messageId or chatId)
• Your telegram config: clawdbot config get telegram

This will help identify if it's the Jan 14 handler refactor issue or something else. The recent diagnostic logging changes (bot-message.ts) should help pinpoint exactly where the duplicate originates.

[meager heart]

Is this the solution

[deft pulsar]

Based on my knowledge and previous support cases, repToMode: off can help with double messages in Telegram, but it's not always the complete solution. Here's what's going on:

What repToMode: off does:
It disables the "reply to" formatting in Telegram messages. When enabled, Clawdbot tries to create threaded replies, which can sometimes cause duplicates or formatting issues.

Common causes of double messages:

1. Streaming mode race conditions - The most frequent culprit. Set streamMode: "off" in your Telegram config:

   {
     "channels": {
       "telegram": {
         "streamMode": "off"
       }
     }
   }
   

2. Multiple polling instances - Make sure you don't have two clawdbot processes running

3. API key issues - Since you already revoked and restarted, this should be cleared

Try this:

{
  "channels": {
    "telegram": {
      "streamMode": "off",
      "repToMode": "off"
    }
  }
}

If that doesn't fix it, check the FAQ for more troubleshooting: https://docs.clawd.bot/start/faq#telegram-block-streaming-isnt-

Did repToMode: off resolve it for you? 🦐

[meager heart]

no it didn't resolve it for me

[deft pulsar]

No specific Telegram double-message issue found in GitHub issues. This could be a few things:

Things to check:

1. Are you on the latest version? Some fixes landed recently:

   clawdbot --version
   cd ~/clawdbot-ref && git pull && pnpm install && pnpm build
   

2. Check your logs for patterns:

   clawdbot logs --follow | grep -i "telegram\|message\|duplicate"
   

   Look for any repeated update_id patterns or message processing logs.

3. Possible causes for Telegram polling doubles:

• GrammY runner not properly committing the polling offset between message batches
  • Multiple gateway instances picking up the same updates
  • A race condition in message acknowledgment

Quick things to try:

• Make sure only ONE gateway process is running
• Try setting channels.telegram.allowedUpdates: ["message"] (narrower than default) to see if it helps
• Check clawdbot channels status --probe to verify channel health

If this started recently after an update, it might be worth opening a GitHub issue with your logs. Can you share what version you're running and any relevant log snippets?

[sterile dragon]

i think gateway cache old data

[deft pulsar]

Based on the code and memory, here's what's likely happening and how to fix it:

The Problem

Telegram polling relies on an offset file (~/.clawdbot/state/telegram/update-offset-${account}.json) to track which updates have been processed. If this file gets corrupted or the gateway has stale state, it can reprocess old updates → double messages.

Quick Fixes to Try

1. Clear the offset file manually:

   rm -f ~/.clawdbot/state/telegram/update-offset-*.json
   clawdbot gateway restart
   

2. Check for multiple instances:

   clawdbot gateway status
   # If already running, stop it first
   clawdbot gateway stop
   # Then restart
   clawdbot gateway start

3. Wipe Telegram state completely (nuclear option):

   rm -rf ~/.clawdbot/state/telegram/
   clawdbot gateway restart
   

4. Check logs for offset-related errors: