Help center
Short, practical guides to getting real work done with AgentMesh. I built the product and wrote these against the live version, so every click below is one I have actually done.
Where to start
| If you… | Read this |
|---|---|
| You have never run an agent | Get started in 3 steps |
| You want to be walked through a whole build | Build a support agent, step by step |
| You want an agent to act on your own systems | Give agents tools & integrations |
| You want it to run without you | Automate: schedules & webhook triggers |
| A run finished and something is wrong | How to troubleshoot a failed run |
Frequently asked questions
Do I need to pay to try this?
No. The free plan needs no card. Runs on it are billed in tokens against a daily allowance, and a worked support-ticket example comes to $0.0144 of usage. Pro is $29 a month when you need the higher limits.
Do I need to write code?
No. An agent is a name, a set of instructions in plain English, and optionally some tools it is allowed to call. The three-step guide below builds a working one in about two minutes.
How do I know what an agent actually did?
Every run records each step, what it received, what it returned, its tokens and its cost — including the steps that failed. Source: one real run and one real failure, published in full so you can read them before signing up.
Also: Compare AgentMesh · AI agents by industry
All guides
Basics
- Get started in 3 steps
From sign-up to your first working agent.
- Templates, plans & billing
Start from something that works, and what each plan actually allows.
- Billing, plans & the Pro trial
Plans, the 10-day trial, invoices, and how to cancel.
Build guides
- Build a support agent that answers a refund ticket
Ten minutes, every field named, one real tool, and the run read line by line.
- Build a client job you can hand over — for consultants
Qualify inbound for a client, and end with a receipt you can send them.
- Move one judgment step off your own server — for n8n self-hosters
Keep n8n for the wiring. Hand one step that needs judgment to an agent.
- Replace one per-task step and measure the bill
For Zapier and Make users: move one step, then compare the two bills honestly.
Agents
- Create and configure an agent
Name it, write its instructions in the Configuration box, activate it.
- Give agents tools & integrations
Let agents call APIs, email, Slack, Discord, Telegram, Notion and your calendar.
Workflows
- Build a multi-agent workflow
Make several agents hand work to each other.
- How to edit an existing workflow
Change it, save a new version, and prove the change took effect.
Automation
Concepts
- What is an AI agent?
A program that reasons about a task and then acts on it, using tools.
- AI agents vs automation tools: what’s the difference?
Deterministic plumbing versus per-case judgment, and when each is wrong.
- What can AI agents do for a small business?
The jobs agents are actually good at, what they cost, and what to keep away from them.
- What is multi-agent orchestration?
Several focused agents coordinated, instead of one prompt doing everything.
- What is an AI workflow?
Several agents and steps run in a defined order, logged and repeatable.
How-to
- How to automate customer support with AI agents
Triage, draft, escalate — and keep a human on the send button.
- How to run AI agents automatically (schedules & webhooks)
On a timetable, or the moment an event fires. Plus what unattended failure looks like.
- How to automate appointment scheduling with AI agents
Read the request, check the calendar, create the event — via Google Calendar.
- How to automate lead qualification with AI agents
Score each inbound lead, summarise it, and route it while it is warm.
- How to connect AI agents to your APIs
The http_request tool: auth headers, response budgets, and the redirect rules.
- How to troubleshoot a failed run
Read the step rows, not the status. Fix the cause, then Replay.
- How to use and customize a marketplace template
Use Template makes an agent. Fork makes your own copy to edit.