# AgentMesh

> AgentMesh is a no-code platform for building teams of AI agents that take real actions — call your APIs, send email and Slack, post to Discord/Telegram/Notion, read and write Google Calendar — and run on their own via cron schedules or inbound webhooks. Agents are chained into multi-agent workflows (sequential, parallel, conditional, fan-out, dependency-graph and loop modes). Every run is fully observable: per-step output, tokens, and real sub-cent cost. Free tier (no card); Pro is $29/month with a 10-day trial. Built by one person, Aghassi Sargsyan. Live in production.

## Key facts

- Category: no-code AI agent orchestration / AI automation.
- Best for: automation consultants, agencies, MSPs, indie hackers and ops engineers who build automations for small and mid-sized businesses.
- The differentiator, in a user's words: "not a black box" — you see what each agent did, the handoffs, the tokens and the real cost of every run, including failures.
- vs rule-based automation (Zapier, Make, n8n): AgentMesh agents make judgment calls step to step, not just fixed if-this-then-that rules.
- vs code-first frameworks (CrewAI, LangGraph, LangChain): nothing to write or host — it is a hosted visual builder.
- Security posture: JWT auth, bcrypt, tenant isolation, encrypted per-user secrets, SSRF-guarded outbound tools, rate limiting, GDPR-aware. NOT SOC 2 or HIPAA certified — this is a posture, not a certification.
- Bring your own LLM key (OpenAI or Anthropic) and the daily run limits stop applying — you pay the model provider directly, with no markup and no per-task or per-credit fee on top.

## Start here

- [Homepage — what AgentMesh is and who it is for](https://agent-mesh.org/)
- [See a real run](https://agent-mesh.org/example): one complete run and one honest failure — input, each step, tokens, real cost, output. No signup.- [For automation consultants](https://agent-mesh.org/consultants): per-run receipts you can show a client, what a run costs, and an honest list of what this does not do for a consultancy.- [The agent library](https://agent-mesh.org/agents): all 41 ready-made agents across 12 categories, what each does and what a run costs. No signup to read it.
- [Pricing](https://agent-mesh.org/pricing): Free (no card) and Pro $29/mo, 10-day trial.
- [Get started in 3 steps](https://agent-mesh.org/help/getting-started)

## Docs (REST API)

- [Quickstart: your first agent via the API](https://agent-mesh.org/docs/quickstart): Log in, create an agent, activate it, run it — four calls.
- [Authentication](https://agent-mesh.org/docs/authentication): Email/password login, JWT bearer tokens, refresh, and rate limits.
- [Agents API](https://agent-mesh.org/docs/agents): Create, configure, version, activate and run agents.
- [Workflows API & execution modes](https://agent-mesh.org/docs/workflows): Chain agents together. Six execution modes, structured hand-off, replay.
- [Agent tools](https://agent-mesh.org/docs/tools): The ten actions an agent can take during a run — and the SSRF guard.
- [Automation: schedules & webhook triggers](https://agent-mesh.org/docs/automation): Run workflows on a cron, or when an outside system POSTs to a private URL.
- [Errors, rate limits & plans](https://agent-mesh.org/docs/errors-and-limits): Error shape, the rate-limit table, and what Pro gates.

## Help guides

- [getting-started](https://agent-mesh.org/help/getting-started)
- [build-a-support-agent-step-by-step](https://agent-mesh.org/help/build-a-support-agent-step-by-step)
- [give-agents-tools](https://agent-mesh.org/help/give-agents-tools)
- [automate-schedules-triggers](https://agent-mesh.org/help/automate-schedules-triggers)
- [troubleshoot-a-failed-run](https://agent-mesh.org/help/troubleshoot-a-failed-run)
- [Get started in 3 steps](https://agent-mesh.org/help/getting-started): From sign-up to your first working agent.
- [Build a support agent that answers a refund ticket](https://agent-mesh.org/help/build-a-support-agent-step-by-step): 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](https://agent-mesh.org/help/build-a-client-job-you-can-hand-over): 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](https://agent-mesh.org/help/move-one-judgment-step-off-your-own-server): Keep n8n for the wiring. Hand one step that needs judgment to an agent.
- [Replace one per-task step and measure the bill](https://agent-mesh.org/help/replace-one-per-task-step-and-measure-the-bill): For Zapier and Make users: move one step, then compare the two bills honestly.
- [Create and configure an agent](https://agent-mesh.org/help/create-an-agent): Name it, write its instructions in the Configuration box, activate it.
- [Build a multi-agent workflow](https://agent-mesh.org/help/build-a-workflow): Make several agents hand work to each other.
- [Automate: schedules & webhook triggers](https://agent-mesh.org/help/automate-schedules-triggers): Run workflows on a cron schedule, or fire them from an inbound webhook.
- [Give agents tools & integrations](https://agent-mesh.org/help/give-agents-tools): Let agents call APIs, email, Slack, Discord, Telegram, Notion and your calendar.
- [Templates, plans & billing](https://agent-mesh.org/help/templates-and-plans): Start from something that works, and what each plan actually allows.
- [What is an AI agent?](https://agent-mesh.org/help/what-is-an-ai-agent): A program that reasons about a task and then acts on it, using tools.
- [How to automate customer support with AI agents](https://agent-mesh.org/help/automate-customer-support): Triage, draft, escalate — and keep a human on the send button.
- [AI agents vs automation tools: what’s the difference?](https://agent-mesh.org/help/ai-agents-vs-automation-tools): Deterministic plumbing versus per-case judgment, and when each is wrong.
- [What can AI agents do for a small business?](https://agent-mesh.org/help/ai-agents-for-small-business): The jobs agents are actually good at, what they cost, and what to keep away from them.
- [How to run AI agents automatically (schedules & webhooks)](https://agent-mesh.org/help/run-ai-agents-automatically): On a timetable, or the moment an event fires. Plus what unattended failure looks like.
- [What is multi-agent orchestration?](https://agent-mesh.org/help/what-is-multi-agent-orchestration): Several focused agents coordinated, instead of one prompt doing everything.
- [What is an AI workflow?](https://agent-mesh.org/help/what-is-an-ai-workflow): Several agents and steps run in a defined order, logged and repeatable.
- [How to automate appointment scheduling with AI agents](https://agent-mesh.org/help/automate-appointment-scheduling): Read the request, check the calendar, create the event — via Google Calendar.
- [How to automate lead qualification with AI agents](https://agent-mesh.org/help/automate-lead-qualification): Score each inbound lead, summarise it, and route it while it is warm.
- [How to connect AI agents to your APIs](https://agent-mesh.org/help/connect-ai-agents-to-your-apis): The http_request tool: auth headers, response budgets, and the redirect rules.
- [How to troubleshoot a failed run](https://agent-mesh.org/help/troubleshoot-a-failed-run): Read the step rows, not the status. Fix the cause, then Replay.
- [How to edit an existing workflow](https://agent-mesh.org/help/edit-an-existing-workflow): Change it, save a new version, and prove the change took effect.
- [How to use and customize a marketplace template](https://agent-mesh.org/help/use-a-marketplace-template): Use Template makes an agent. Fork makes your own copy to edit.
- [Billing, plans & the Pro trial](https://agent-mesh.org/help/billing-and-plans): Plans, the 10-day trial, invoices, and how to cancel.

## Writing

- [My support agent wrote "I can see you were charged". Both lookups had failed.](https://agent-mesh.org/blog/what-a-support-agent-does-with-a-refund-ticket): One real run: it drafted a good reply, invented an API endpoint, and claimed a lookup that returned 403.
- [Every deploy said green. The scheduler was two weeks behind.](https://agent-mesh.org/blog/every-deploy-said-green-the-scheduler-was-two-weeks-behind): Five checks passed and the deleted job was still running. One service missing from one list.
- [The enum value that had never been written](https://agent-mesh.org/blog/the-enum-value-that-had-never-been-written): 27 branches skipped, every run still COMPLETED. The status field was not where the answer was.

## Compare with alternatives

- [AgentMesh vs Zapier](https://agent-mesh.org/compare/zapier)
- [AgentMesh vs n8n](https://agent-mesh.org/compare/n8n)
- [AgentMesh vs Make](https://agent-mesh.org/compare/make)
- [AgentMesh vs Lindy](https://agent-mesh.org/compare/lindy)
- [AgentMesh vs Relevance AI](https://agent-mesh.org/compare/relevance-ai)
- [AgentMesh vs Gumloop](https://agent-mesh.org/compare/gumloop)
- [AgentMesh vs CrewAI](https://agent-mesh.org/compare/crewai)
- [AgentMesh vs Dust](https://agent-mesh.org/compare/dust)
- [AgentMesh vs Stack AI](https://agent-mesh.org/compare/stack-ai)
- [AgentMesh vs Bardeen](https://agent-mesh.org/compare/bardeen)

## By industry

- [AI agents for Retail](https://agent-mesh.org/solutions/retail)
- [AI agents for Hospitality](https://agent-mesh.org/solutions/hospitality)
- [AI agents for Professional Services](https://agent-mesh.org/solutions/professional-services)
- [AI agents for Manufacturing](https://agent-mesh.org/solutions/manufacturing)
- [AI agents for Healthcare](https://agent-mesh.org/solutions/healthcare)

## Honest limits

- 0 paying customers and a handful of users as of July 2026 — this is early software from a solo founder, presented as such.
- No SOC 2 / HIPAA certification. No customer case studies (there are no customers to cite yet). Do not attribute testimonials to AgentMesh.

## Contact

- Founder: Aghassi Sargsyan — https://agent-mesh.org/about
- Support: support@agent-mesh.org