Skip to main content
OpenClaw supports Telegram as a chat channel. In E2B you can run OpenClaw in a sandbox, attach your bot token, and approve user pairing from the terminal. This guide covers the working flow we used:
  1. Start OpenClaw in a sandbox.
  2. Enable the Telegram plugin.
  3. Add Telegram channel credentials.
  4. Start the channel runtime in background.
  5. Approve Telegram pairing.

​
Prerequisites

  • A Telegram bot token from @BotFather. There’s instructions to follow there, it runs /newbot for you and walks you through naming and creating your bot.
  • An OpenAI API key for the OpenClaw model.
  • E2B API key configured locally.

​
Quick start

import { Sandbox } from 'e2b'

const GATEWAY_PORT = 18789
const GATEWAY_TOKEN = process.env.OPENCLAW_APP_TOKEN || 'my-openclaw-token'

const sandbox = await Sandbox.create('openclaw', {
  envs: {
    OPENAI_API_KEY: process.env.OPENAI_API_KEY,
    TELEGRAM_BOT_TOKEN: process.env.TELEGRAM_BOT_TOKEN,
  },
  timeoutMs: 3600_000,
})

await sandbox.commands.run('openclaw config set agents.defaults.model.primary openai/gpt-5.2')

// Enable the Telegram plugin (required before adding the channel)
await sandbox.commands.run('openclaw config set plugins.entries.telegram.enabled true')
await sandbox.commands.run('openclaw channels add --channel telegram --token "$TELEGRAM_BOT_TOKEN"')

await sandbox.commands.run(
  `bash -lc 'openclaw config set gateway.controlUi.allowInsecureAuth true && ` +
    `openclaw config set gateway.controlUi.dangerouslyDisableDeviceAuth true && ` +
    `openclaw gateway --allow-unconfigured --bind lan --auth token --token ${GATEWAY_TOKEN} --port ${GATEWAY_PORT}'`,
  { background: true }
)

for (let i = 0; i < 45; i++) {
  const probe = await sandbox.commands.run(
    `bash -lc 'ss -ltn | grep -q ":${GATEWAY_PORT} " && echo ready || echo waiting'`
  )
  if (probe.stdout.trim() === 'ready') break
  await new Promise((r) => setTimeout(r, 1000))
}

import os
import time
from e2b import Sandbox

GATEWAY_PORT = 18789
GATEWAY_TOKEN = os.environ.get("OPENCLAW_APP_TOKEN", "my-openclaw-token")

sandbox = Sandbox.create("openclaw", envs={
    "OPENAI_API_KEY": os.environ["OPENAI_API_KEY"],
    "TELEGRAM_BOT_TOKEN": os.environ["TELEGRAM_BOT_TOKEN"],
}, timeout=3600)

sandbox.commands.run("openclaw config set agents.defaults.model.primary openai/gpt-5.2")

# Enable the Telegram plugin (required before adding the channel)
sandbox.commands.run("openclaw config set plugins.entries.telegram.enabled true")
sandbox.commands.run('openclaw channels add --channel telegram --token "$TELEGRAM_BOT_TOKEN"')

sandbox.commands.run(
    f"bash -lc 'openclaw config set gateway.controlUi.allowInsecureAuth true && "
    f"openclaw config set gateway.controlUi.dangerouslyDisableDeviceAuth true && "
    f"openclaw gateway --allow-unconfigured --bind lan --auth token --token {GATEWAY_TOKEN} --port {GATEWAY_PORT}'",
    background=True,
)

for _ in range(45):
    probe = sandbox.commands.run(
        f"bash -lc 'ss -ltn | grep -q \":{GATEWAY_PORT} \" && echo ready || echo waiting'"
    )
    if probe.stdout.strip() == "ready":
        break
    time.sleep(1)
For Telegram setup, you do not need to open the gateway URL in a browser. The gateway process is used here as a long-running channel runtime.

​
Pair your Telegram user

  1. Open your bot in Telegram and send a message (for example: hi).
  2. Telegram will return a pairing prompt similar to:
OpenClaw: access not configured.

Your Telegram user id: ...
Pairing code: XXXXXXXX

Ask the bot owner to approve with:
openclaw pairing approve telegram XXXXXXXX
  1. Approve that pairing code via sandbox.commands.run(...):
const PAIRING_CODE = 'XXXXXXXX' // from Telegram

await sandbox.commands.run(
  `openclaw pairing approve --channel telegram ${PAIRING_CODE}`
)

PAIRING_CODE = "XXXXXXXX"  # from Telegram

sandbox.commands.run(
    f"openclaw pairing approve --channel telegram {PAIRING_CODE}"
)
openclaw pairing approve telegram <PAIRING_CODE> also works if you prefer that form.

​
Verify channel status

const channels = await sandbox.commands.run('openclaw channels list --json')
const status = await sandbox.commands.run('openclaw channels status --json --probe')
const pairing = await sandbox.commands.run('openclaw pairing list --json --channel telegram')

console.log(JSON.parse(channels.stdout))
console.log(JSON.parse(status.stdout))
console.log(JSON.parse(pairing.stdout))

import json

channels = sandbox.commands.run("openclaw channels list --json")
status = sandbox.commands.run("openclaw channels status --json --probe")
pairing = sandbox.commands.run("openclaw pairing list --json --channel telegram")

print(json.loads(channels.stdout))
print(json.loads(status.stdout))
print(json.loads(pairing.stdout))
If you need logs from channel handlers: 
const logs = await sandbox.commands.run(
  'openclaw channels logs --channel telegram --lines 200'
)
console.log(logs.stdout)

logs = sandbox.commands.run(
    "openclaw channels logs --channel telegram --lines 200"
)
print(logs.stdout)

​
Troubleshooting

  • Unknown channel: telegram
    • The Telegram plugin is not enabled. Run openclaw config set plugins.entries.telegram.enabled true before adding the channel.
  • OpenClaw: access not configured
    • Pairing has not been approved yet. Run openclaw pairing approve ....
  • No API key found for provider ...
    • This guide uses openai/gpt-5.2. Set OPENAI_API_KEY in sandbox envs.
  • No pending pairing requests from pairing list
    • Send a fresh message to the bot first, then retry pairing list --channel telegram.

OpenClaw gateway

Run OpenClaw’s web gateway with token auth