Industries
Manage your AI chatbot from Claude Code
Ask in plain English. Your AI agent reads your chatbot's real analytics, finds the questions it fails to answer, and fixes them. It confirms every change before making it.
Two commands in Claude Code. Works on the free plan.
Trusted by 1,000+ businesses across SaaS, e-commerce, and agencies
What you can ask
Say what you want. It works out the steps.
"Why is my bot failing to answer questions?"
It reads your analytics, groups the questions your chatbot missed, and shows you which content is missing.
"Add our new pricing page and retrain"
Adds the page as a knowledge source and retrains, so your chatbot answers from it straight away.
"Hand off to a human after two failed answers"
Sets up escalation so a person takes over when your chatbot cannot help, with email alerts to your team.
"Move this bot to a cheaper model"
Switches between OpenAI, Anthropic, Google, and xAI models your plan allows. Bring your own API key if you prefer.
"Which pages should I add next?"
Compares what visitors ask against what your chatbot was trained on, and tells you what to publish.
"Show me this week's conversations"
Reads your inbox, summarises what people asked, marks threads as done, and pulls the leads your chatbot captured.
Turn conversations into leads
Optimize for lead capture in one sentence
Your agent checks how many conversations turn into leads, then turns on lead capture, sets up human handoff, and adds suggested messages that start the conversation. You approve the plan before anything changes.
Clear the inbox
Triage support without leaving your terminal
Read open conversations, see what people asked, and mark the handled ones as done. Your agent points you at the threads that still need a person.
It asks before it changes anything
Your AI agent reads freely, but every change is confirmed with you first. It can only reach the chatbots your API token allows, and you can delete that token at any time.
Compatibility
Works with any MCP client
Claude Code installs in two commands. For Cursor, VS Code, or any other client that speaks the Model Context Protocol, point it at our server and add your API token.
Create your token in the dashboard under AI Agents.
Run your chatbot from your terminal
Create a free SiteSpeakAI account, install the plugin, and ask your first question in under five minutes.