AgentMesh
No-code AI agent teams that run your operations
Start new thread

AgentMesh - No-code AI agent teams that run your operations

by•
AgentMesh lets you build teams of AI agents that actually do things — answer customers, call your APIs, send email and Slack, update your calendar — and hand off to each other in workflows. No code. Free to start; Pro is $29/mo.

Add a comment

Replies

Best
Maker
📌
Six months ago I was doing a course on LLM engineering. Today the thing I built from it is live, in production, taking payments. Most "AI agent" tools stop at chat. I wanted agents that take real actions and run on their own. So AgentMesh agents call any API, send email/Slack/Discord/Telegram, write to Notion, fire webhooks, and update Google Calendar — and you orchestrate several of them together (sequential, parallel, conditional, fan-out, DAG, loop) on a no-code canvas. They run on a cron schedule or when a webhook fires, and you can replay any run in one click. It's built solo — 546 commits, 486 backend tests — and it's honest about what it isn't: it's a monolith, not microservices, and it's GDPR-aware but not SOC 2 or HIPAA certified. I say so on the site. One click on the dashboard spins up a starter agent and runs it, so you can see it work in seconds. Free to build; Pro ($29/mo) unlocks every orchestration mode. Would love your feedback — especially what you'd automate first. 🙏

Love that the agents can actually hand off work to each other without me babysitting them. Set up a customer question → Slack ping → calendar booking flow in like ten minutes.

Maker

 This made my day 🙏 That hand-off — one agent's output becoming the next one's input — is the whole reason I built it. "Without me

babysitting them" is exactly the feeling I was going for. What was the customer-question part reading from — email, a form, a webhook?

Congrats on the launch! Looks powerful. Which integration is most popular so far?

Maker

 Thanks Ghevond! 🙏 Slack is the most-used so far — people love an agent that reads a message, decides, and pings the right channel. Google

Calendar and plain HTTP (call any API) are close behind. What would you want to connect first?

Would love to see a built-in cost dashboard showing token spend per agent and per workflow run — helps teams track ROI and catch runaway loops before the bill surprises them.

Maker

 This is a great ask — and most of it already exists, so I should surface it better. There's a Cost Analytics view (Analytics → Costs) with total spend, a daily cost trend, avg per execution, and cost by agent. Token usage is tracked per run too.

Would love to see a visual debug mode that shows the handoffs between agents in real time, like a flowchart of which agent triggered which and why. Right now if a workflow stalls I'm guessing where it broke. That kind of observability would make AgentMesh way more trustworthy for real production use.

Maker

 This is the sharpest ask I've had today, and you're right about the gap.

What's there now: every run has a Step Executions view — each step with its status, duration, and full input/output, plus timestamped logs and one-click replay. So you can see which step failed and what it received. But it's a list, not a picture.

What's missing is exactly what you're describing: the canvas lighting up live — this step is running, this one handed off to that one, here's why it branched. The canvas today is build-time only. That's a fair miss, and "observability makes it trustworthy for production" is exactly the right

framing.

Adding it to the roadmap. Would live step highlighting on the canvas + the branch reason ("took this path because category == billing") cover it, or would you want a full timeline/trace view too?

the handoff workflow between agents looks really clean, like you can actually see what each one is doing at any moment instead of it being a black box

Maker

 Thank you 🙏 "Not a black box" is exactly what I was going for — every run shows each step, what it received, what it returned, and how long it took. Watching agents hand off with no idea what happened in between was the thing that annoyed me most about other tools.

What are you thinking of pointing it at?