← Tendabot MCP
Manage bots

Edit your AI agent by asking Claude.

Six tools cover the whole lifecycle of an agent: find it, read it, change it, and check the change before a customer sees it.

Things you can just ask

Real prompts, not pseudo-code. Claude picks the tools itself.

List all my bots and tell me which one has the most open conversations.
Read the instructions for the store bot and tell me what it would say if someone asked for a refund.
Change the support bot to a cheaper model for order-status questions.
Set office hours to Monday–Friday 9–5 and turn handoff off outside those hours.
Ask the bot "do you ship to Ireland" and show me exactly what it replies.

The 6 tools behind it

ToolNeedsWhat it does
list_botsreadEvery bot across every workspace you administer, with the workspace it belongs to.
get_botreadOne bot in full — instructions, model, settings.
update_bot_instructionswriteRewrite what the bot is told to do. Snapshotted, so you can roll back.
update_bot_settingswriteModel, office hours, handoff, automations.
test_botwriteAsk the bot a question and read the answer, without touching a real conversation.
get_overviewreadConversation counts by stage, per bot, across every workspace.

read works with any key. write is refused unless the key allows it — the server checks before the tool runs, so a read-only key genuinely cannot change anything.

What it will not do

Worth knowing before you wire it into something that matters.

  • Creating a brand-new bot still happens in the console — MCP edits agents that already exist.
  • update_bot_instructions rewrites the whole prompt rather than patching a line. Read it first with get_bot.
  • test_bot needs a write key even though it changes nothing, because it spends credits to generate a reply.

Also over MCP

Ready when you are

Your first agent, answering in five minutes.