> ## Documentation Index > Fetch the complete documentation index at: https://docs.suji.fr/llms.txt > Use this file to discover all available pages before exploring further. # OpenClaw > Multi-channel AI messaging bot. Install, configure, connect channels, and fix the common first-connect issues. [OpenClaw](https://github.com/openclaw/openclaw) is an open-source multi-channel messaging bot. It connects to Telegram, Discord, WhatsApp Cloud API, and Slack, and replies using your choice of AI provider: Anthropic, OpenAI, Google, Mistral, Groq, OpenRouter, xAI, Moonshot, or any OpenAI-compatible endpoint (vLLM, Ollama, LM Studio, LiteLLM, and the like). This page covers running OpenClaw on Suji end-to-end. OpenClaw itself is maintained upstream; Suji provides the [marketplace packaging](https://github.com/suji-hq/suji-templates/tree/main/openclaw). *** ## Before you install You'll need: * **An AI provider API key** (optional — OpenClaw boots unconfigured, so you can also skip this and set it later from the **Terminal** with the `openclaw` CLI). For example: * Anthropic — [console.anthropic.com](https://console.anthropic.com) * OpenAI — [platform.openai.com](https://platform.openai.com) * Or any other supported provider / OpenAI-compatible endpoint. * **At least one channel token** for the platforms you want OpenClaw to respond on. You can start with one and add others later. | Channel | Where to get the token | | ------------------ | ----------------------------------------------------------------------- | | Telegram | [@BotFather](https://t.me/BotFather) | | Discord | [Discord developer portal](https://discord.com/developers/applications) | | WhatsApp Cloud API | Meta Business | | Slack | Your Slack app's config page | *** ## Install Dashboard → **Apps** → **OpenClaw** → **Install**, then pick the VM to install on. | Field | Required | Notes | | ------------------------ | --------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------- | | AI provider | yes | Anthropic, OpenAI, Google, Mistral, Groq, OpenRouter, xAI, Moonshot, or OpenAI-compatible. | | Custom endpoint base URL | only for OpenAI-compatible | e.g. `https://llm.example.com/v1`. | | AI API key | for hosted providers | Encrypted at rest. Optional for OpenAI-compatible endpoints that need no auth. Use **Test connection** to verify the key and model. | | Model | for OpenRouter / Moonshot / OpenAI-compatible | e.g. `gpt-5.5`, `claude-sonnet-4-6`. **Test connection** lists the provider's models, or type any value. | | Gateway access token | no | Unlocks the Control UI. Auto-generated if blank; reveal it later in the Variables tab. | | Telegram bot token | no | Provide to enable Telegram; leave blank to skip. | | Discord bot token | no | Provide to enable Discord; leave blank to skip. | OpenClaw boots with `--allow-unconfigured`, so the AI fields are optional. You can choose **Skip & configure later**, then set the provider, key, and model after install from the **Terminal** tab (pick OpenClaw) using the `openclaw` CLI. See [Configure after a skip install](#configure-after-a-skip-install) below. Recommended VM size: **Small** or larger. (Small = 2 vCPU / 4 GB is the floor.) Click **Deploy**. The install reaches `running` in \~1 minute. You'll see a public URL like `https://.suji.fr` — that's the Control UI. *** ## First connection (two one-time steps) OpenClaw protects the Control UI with two layers. Both run automatically from `localhost`, but over the public tunnel you do them once: 1. Sign in with your **gateway access token**. 2. **Approve** this browser (a one-time device pairing). The install's **Deployment** tab also walks you through these steps with the token revealable inline. ### Step 1 — sign in with the access token Click **Open** on the install page to load the Control UI. Its login screen asks for the gateway access token. That token is the **Gateway access token** from the install form (auto-generated if you left it blank). To get it: * Install detail page → **Deployment** tab → **Getting started** — reveal and copy it there, **or** * Install detail page → **Variables** tab → reveal `OPENCLAW_GATEWAY_PASSWORD`. Paste it into the login screen. It's the same value every time you need it again. ### Step 2 — approve this browser The first time you sign in, OpenClaw asks you to approve this browser and shows a command like: ``` openclaw devices approve [id] ``` Run that exact command (with the real id it showed you) from the dashboard: 1. Instance page → **Terminal** tab → pick **OpenClaw** from the selector. 2. Paste and run the command. 3. Sign in again. The Control UI connects. The approval request is short-lived, so approve it while the sign-in screen is still open. Pairing is **per browser/device**, so you'll repeat this once for each new browser. *** ## Configure after a skip install If you chose **Skip & configure later**, OpenClaw is running but unconfigured: it has no AI provider yet. You set this up from the **Terminal** with the `openclaw` CLI, not from OpenClaw's web UI. Instance page → **Terminal** tab → pick **OpenClaw**, then run the guided setup: ```bash theme={"dark"} openclaw configure ``` It walks you through credentials (provider + API key), the default model, and channels, and applies them to the running gateway. To jump straight to one part, target a section: ```bash theme={"dark"} openclaw configure --section model # provider + model openclaw configure --section channels # Telegram / Discord / WhatsApp / Slack ``` Prefer non-interactive commands? They map one-to-one: ```bash theme={"dark"} openclaw models set anthropic/claude-sonnet-4-6 # / openclaw channels add # add a messaging channel openclaw config get # inspect a setting openclaw config set # change a setting ``` CLI changes write to OpenClaw's persisted config and take effect immediately, so there's no redeploy. A **Restart** re-applies the install form, which re-seeds a default model for your provider. If you set the model from the CLI and later restart, re-run `openclaw models set …`, or pin it once in the **Variables** tab (`AI_MODEL`) so it survives restarts. *** ## Connecting channels You have two ways to add a channel: * **Bot token (Telegram / Discord):** set `TELEGRAM_BOT_TOKEN` or `DISCORD_BOT_TOKEN` in the install's **Variables** tab and **Restart**. Presence of the token enables the channel at startup; there is no separate selector. * **CLI (any channel):** from the **Terminal** (pick OpenClaw), run `openclaw channels add` and follow the prompts. This is how you set up **WhatsApp** (scan the QR code with your phone) and **Slack** (needs both a bot and an app token), which don't fit a single install-time field. For webhook-based channels (WhatsApp Cloud API, Slack events), point the webhook URL in the provider's console at `https://.suji.fr/webhooks/`. Confirm by sending a test message to your bot. *** ## Day-to-day management | Want to… | Where | | ------------------------------------------- | ----------------------------------------------------------------- | | Change AI provider / model / channels | Terminal (pick OpenClaw) → `openclaw configure` | | View live logs | Dashboard → Logs (pick OpenClaw) | | Open a shell as `node` inside the container | Dashboard → Terminal (pick OpenClaw) | | Browse data files | Dashboard → Files (pick the OpenClaw volume) | | Edit `openclaw.json` | Files → open → `Cmd-S` to save → restart from install detail page | | Restart | Install detail page → Restart | | Upgrade to a newer version | Install detail page → Upgrade (when available) | | Remove the install + its data | Install detail page → Uninstall | *** ## Troubleshooting Paste the **Gateway access token** (`OPENCLAW_GATEWAY_PASSWORD`). Reveal it on the install's **Deployment** tab → **Getting started**, or the **Variables** tab. It's the same value every time. The browser isn't paired yet. Open the instance **Terminal** tab (pick OpenClaw), run the `openclaw devices approve [id]` command OpenClaw showed you, then sign in again. Pairing is per browser, so repeat on every new device. OpenClaw isn't reachable through the tunnel. Most common causes: * Container still starting — wait 30 s and retry, especially right after an upgrade. * Container crashing — Logs tab will show the reason. If it's a permission error on `/home/node/.openclaw`, that's a packaging issue — file a Suji support ticket. OpenClaw runs as the `node` user. If the data volume's ownership has drifted, OpenClaw can't write its config. Fix from the host terminal: ```bash theme={"dark"} docker exec --user 0 $(docker ps -q -f label=com.docker.compose.project=) \ chown -R node:node /home/node/.openclaw docker compose -f /etc/suji/installs//compose.yaml restart openclaw ``` Fresh installs handle this automatically via an init-permissions container. The underlying network blocks outbound SMTP on ports 25 and 465. Use port 587 with STARTTLS, or an HTTP-based provider (Resend / Postmark / Mailgun / SendGrid). A firewall rule won't help — the block is upstream of the firewall. Expected. Pairing is per device — approve the new browser's pairing request the same way as the first. *** ## Where things live | What | Inside container | On host | | -------------------------------- | ---------------------- | ----------------------------------------------- | | sqlite, sessions, gateway config | `/home/node/.openclaw` | Named volume `_openclaw-data` | | Logs | stdout/stderr | — | | Compose file (rendered) | — | `/etc/suji/installs//compose.yaml` | | Env (rendered secrets) | injected as env vars | `/etc/suji/installs//.env` (`0600`) | Uninstalling deletes the volume by default. Choose **keep data** during uninstall to preserve it (you can reconnect to it later over SSH). *** ## Recommended size * **Small (2 vCPU / 4 GB)** is the floor for stable operation. * The compose caps OpenClaw at 2048 MiB and 1 vCPU. AI work itself runs at the provider (Anthropic / OpenAI) — local CPU is just message routing, channel adapters, and HTTP roundtrips, which is light. The 2 GB ceiling is what actually matters: sustained high message volume can push memory usage if sessions accumulate, which is when an OOM kill would happen. If you run multiple apps on the same VM, factor those caps in — the VM's total CPU/memory is still shared. *** ## Reporting issues | Class | Where | | ------------------------------------------------------------- | --------------------------------------------------------------------------------- | | OpenClaw bug (wrong AI reply, channel logic, etc.) | [openclaw/openclaw issues](https://github.com/openclaw/openclaw/issues) | | Marketplace packaging bug (compose / manifest / install form) | [suji-hq/suji-templates issues](https://github.com/suji-hq/suji-templates/issues) | | Suji platform bug (dashboard, billing, network) | Support ticket from the dashboard |