🚀
Coming from OpenClaw?
Import your config — no Mac Mini required. Same agent capabilities, $5/mo vs ~$600 hardware.
Import Config

Deploy a New Agent

Configure and launch a cloud-hosted AI agent.

Start from a Template (optional)

Pick a use case to pre-fill the system prompt.

1. Agent Identity

Lowercase, no spaces. Used in URLs and the dashboard.

Custom domain for webhook endpoints. +$2/mo add-on

This is the system message sent to the LLM on every conversation. Be specific about tone, scope, and boundaries.

2. LLM Provider
Anthropic
Claude
OpenAI
GPT-4
Google
Gemini
Moonshot
Kimi K2.5
MiniMax
M2.5

Encrypted at rest. Never stored in plain text or logs. Get an Anthropic API key →

0 = deterministic, 1 = creative

3. Integrations

How should users interact with this agent?

Telegram
Respond to messages in a Telegram bot
iUses Telegram's Bot API with long polling. No extra server needed — runs inside your agent container at zero additional cost. Included
Discord
Respond to @mentions and DMs via Gateway WebSocket
iConnects via Discord's Gateway WebSocket. Responds to @mentions in servers and all DMs. Completely free — no API costs. Included
WhatsApp
Respond via WhatsApp Cloud API webhooks
iUses Meta's WhatsApp Cloud API with webhooks. 1,000 free service conversations/month. Costs are billed by Meta, not TamaleBot. Included
Slack
Respond in Slack channels and DMs
iConnects via Slack's Socket Mode (WebSocket). No public endpoint or extra infrastructure required. Runs inside your agent container. Included
Email (SMTP)
Send and receive emails
iSends via your SMTP server (Gmail, SES, etc). No extra infrastructure — your agent connects directly to your mail provider. Included
Webhooks
Trigger agent via HTTP POST, receive callbacks
iYour agent gets a unique HTTPS endpoint. Send a POST request to trigger it, or configure outbound webhooks to notify other services. Included
iMessage
Send and receive iMessages (requires Mac relay)
iiMessage requires macOS hardware — it can't run in Linux containers. We maintain a shared Mac relay fleet. Coming in a future update. Coming Soon
4. Tools & Permissions

What can this agent do? All actions are policy-checked and audit-logged.

Shell Commands
Run terminal commands in the agent container
iExecute bash commands inside a hardened Docker container. All commands pass through the security policy engine before running. Included
File Read / Write
Read and write files in the agent workspace
iRead and write files within the agent's isolated workspace directory. Sensitive paths (SSH keys, credentials) are blocked by the policy engine. Included
Web Browsing (fetch)
Fetch URLs and extract text content
iLightweight HTTP fetch — retrieves pages and extracts text. Good for APIs and static content. No JavaScript rendering. Included
Full Browser (Chromium)
Headless Chrome for JS-rendered pages, screenshots, form fill
iRuns a headless Chromium instance alongside your agent. Needed for JS-rendered pages, screenshots, form filling, and browser automation. Uses ~500MB extra RAM.
Sub-Agents
Let this agent spawn and control separate worker agents
iYour main agent becomes the brain. It can spin up sub-agents in separate containers to handle tasks — research, coding, data processing — then collect results. Each sub-agent is isolated with its own policy.
1
5. Security Policy

Control what the agent is allowed to do. You can customize the policy after deployment.

Default (Recommended)
Blocks destructive commands, sensitive files, and data exfiltration. Good for most use cases.
Strict
Read-only filesystem, no outbound network except allowed domains. For compliance-sensitive environments.
Permissive
Minimal restrictions. Full shell access, outbound network allowed. For development and testing.
Custom YAML
Upload your own policy file or edit after deployment.

Comma-separated. Leave blank to allow all outbound traffic (default policy).

6. Resource Limits

Agent pauses when LLM API spend hits this limit. Change anytime.

Rate limit for inbound messages. Prevents runaway costs.

Writable disk space for the agent container.

6. Access Control (optional)

Restrict SSH and Git access for added security. Leave empty to allow all.

Comma-separated. Empty = allow all.

Comma-separated. Empty = allow all.

Estimated Monthly Cost

Cloud Agent (base) $5.00
Total $5.00 /mo
Billed daily (~$0.17/day). Destroy anytime.

Deploys to Cloudflare Containers. Your agent gets its own isolated VM.

Your API key is billed separately by your LLM provider. TamaleBot only charges for hosting and add-ons — we never mark up API costs.