OpenAI-compatibleMulti-channel

OpenClaw + MegaBrain Gateway

OpenClaw is an open-source AI agent platform that brings conversational AI to messaging channels. Connect it to MegaBrain to access 500+ models across Telegram, Discord, Slack, WhatsApp, Signal, and iMessage — all with one API key.

Supported channels

TelegramDiscordSlackWhatsAppSignaliMessage

Prerequisites

Node.js 18 or later — nodejs.org
A MegaBrain API key from your profile
A bot token for the messaging channel you want to connect (Telegram, Discord, etc.)

Quick setup (one line)

Install OpenClaw, then run a single non-interactive onboard command:

Install OpenClaw

npm install -g openclaw

Onboard with MegaBrain

One command — no prompts:

openclaw onboard --non-interactive \
  --auth-choice custom-api-key \
  --custom-base-url "https://getmegabrain.com/api/gateway/v1" \
  --custom-model-id "anthropic/claude-sonnet-4-6" \
  --custom-api-key "$MEGABRAIN_API_KEY" \
  --custom-compatibility openai

Replace $MEGABRAIN_API_KEY with your key from your profile.

Start the gateway

openclaw gateway run

OpenClaw connects to your messaging channel and routes requests through MegaBrain.

Coming soon

Native MegaBrain provider

We've submitted a PR to OpenClaw to add MegaBrain as a first-class provider. Once merged, onboarding will be a single flag:

openclaw onboard --auth-choice megabrain-api-key --megabrain-api-key "$MEGABRAIN_API_KEY"

Manual configuration

Prefer editing config directly? Use ~/.openclaw/openclaw.json:

{
  "provider": "megabrain",
  "apiKey": "mb-your-key-here",
  "baseURL": "https://getmegabrain.com/api/gateway/v1",
  "model": "anthropic/claude-sonnet-4-6",
  "channels": {
    "telegram": {
      "botToken": "your-telegram-bot-token"
    }
  }
}

Or set via environment variables:

export OPENCLAW_API_KEY="mb-your-key-here"
export OPENCLAW_BASE_URL="https://getmegabrain.com/api/gateway/v1"
export OPENCLAW_MODEL="anthropic/claude-sonnet-4-6"

Recommended models

Pass any MegaBrain model ID as the model value:

# Fast and capable — best for most chats
anthropic/claude-sonnet-4-6

# Most capable — complex analysis and reasoning
anthropic/claude-opus-4-8

# Cost-effective — high volume, simple queries
google/gemini-2.5-flash

# Auto-select the best model per message
auto-balanced

auto-balanced routes each message to the most cost-effective model that can handle it — recommended for bots with mixed query complexity. Browse all models on the models page.

Different models per channel

You can assign a different model to each messaging channel. Useful for running a cheaper model on a high-volume public bot and a flagship model on a private workspace:

{
  "channels": {
    "telegram": {
      "botToken": "...",
      "model": "google/gemini-2.5-flash"
    },
    "slack": {
      "botToken": "...",
      "model": "anthropic/claude-opus-4-8"
    }
  }
}

Fallback model

Configure a fallback model to keep the bot running if the primary model is unavailable:

{
  "model": "anthropic/claude-sonnet-4-6",
  "fallbackModel": "google/gemini-2.5-flash"
}

Troubleshooting

Authentication error / 401

Verify your API key starts with mb- and that baseURL is exactly https://getmegabrain.com/api/gateway/v1 — no trailing slash.

Model not found / 404

Model IDs are case-sensitive and follow the format provider/model-name. Check the models page for the exact ID.

Bot not responding on Telegram

Make sure the bot token is correct and the bot has been started via /start in the Telegram chat. Run openclaw gateway status to check the connection state.

← OpenCode Claude Code →