An AI agent builder where you describe the task
Xagent is an AI agent builder for business teams. You write what you need in plain language. Xagent writes the plan as an execution graph, runs it with your connected tools and hands back the result. Save it as an agent to run again or put on a schedule.
How to build an AI agent in Xagent
Four moves, and none of them is drawing a flowchart.
Describe the task
Write what you need the way you would brief a colleague. The request is the whole spec.
Read the plan
Xagent turns the request into an execution graph with one card for each step and the tools that step needs.
Watch it run
Every tool call and result shows while the task runs, with the token count. You can pause a running task.
Save it as an agent
Keep its role, model, tools and knowledge. Run it again, or start it on a schedule, a webhook or a new Gmail message.
An example run: the Friday status report
The operations lead at Northwind Freight wants the weekly update written without chasing anyone.
- CalendarReads this week's meetingsReturns the titles, times and who attended.
- DriveFinds the notes for each meetingReturns the documents from the team folder.
- Knowledge baseChecks the project pagesReturns the passages that match the meeting topics.
- DocumentsWrites the updateDrafts a one-page status note in your usual headings.
- OutlookDrafts the emailPuts the note in a draft addressed to the team.
An AI agent builder with no workflow to draw
In a canvas-style builder you draw each step and wire the tools together before anything runs. Xagent works the other way round. You say what the finished work should look like, and the plan comes from the request.
When the work changes, you change the request and Xagent writes a new plan. Nobody has to maintain a diagram.
- No diagram to draw before the first run
- The plan is on screen before the tools start
- Every call and result is visible as it happens
A no-code AI agent builder, and where code comes in
Most agents need no code at all. Here is where a developer helps.
Describe and run
Plain-language requests, templates and the native Google, Microsoft, Meta and LinkedIn connectors need no code. Your team can do all of it.
Connect your own systems
Other systems, including internal ones, connect through MCP servers. Someone technical sets each one up once.
Build into your product
The API, the chat widget and the SDK put agents inside your own products. This part is for developers.
Build AI agents from a template
Start from one of eight templates, point it at your knowledge and your tools, then change the request until it fits.
Support chatbot
Answers questions from your knowledge base and drafts replies.
Knowledge baseGmailInbound leads
Reads new enquiries, researches the sender and drafts a reply.
GmailLinkedInLinkedIn post
Turns a brief into a post for your page.
LinkedInDrafts the emails you keep writing from a short request.
GmailOutlookContent
Drafts articles and pages from a brief and your own documents.
DocumentsDriveInbox manager
Sorts a shared inbox so people see the important messages first.
GmailOutlookSlides
Drafts a deck from a request and your source material.
SlidesOneDriveDocuments
Writes reports and notes and saves them where you keep them.
DocumentsDriveFive roles ship in every workspace
You do not start from a blank agent. Each workspace comes with five specialists and a planner to coordinate them.
Support
Answers questions from your knowledge base and drafts replies.
Growth
Researches leads and drafts outreach and posts.
Planner
Breaks a request into steps and hands each to the right agent.
Knowledge
Searches your documents and writes the report.
Studio
Drafts decks, pages and visuals.
Put the agents you build inside your own product
An agent you save can leave the Xagent workspace.
API
Call your agents from your own software.
Chat widget
Embed an agent as a chat widget in your product.
SDK
Make your own app agentic with a few lines of code.
Triggers
Run agents on a schedule, from a webhook or on a new Gmail message.
What the Xagent AI agent builder costs
Pro is US$59 a month for 2 users, 5k actions and triggers a month and 10k AI credits. Team is US$299 a month for 10 users, 25k actions and triggers a month and 50k AI credits. Both include unlimited agents and let you bring your own LLM key. Enterprise is talk to sales, on Xagent cloud or self-hosted. Plan details are on the pricing page.
AI agent builder: frequently asked questions
What is an AI agent builder?
An AI agent builder is software for creating AI agents that do a job with your tools. Some start from a blank canvas where you draw the steps. Xagent starts from a plain-language request and writes the plan for you.
Can I build an AI agent without code?
Yes for most agents. You describe the task, read the plan and save the result. Connecting an internal system through an MCP server, or building agents into your own product with the API and SDK, is work for a developer.
Which models can I use?
Xagent works with open models and API providers, including any OpenAI-compatible endpoint, and with models self-hosted through Xinference. Don't see the model you need? Talk to us about adding it.
Which tools can an agent use?
Native connectors cover Google (Gmail, Calendar, Drive, Maps), Microsoft (Outlook, Teams, OneDrive), Meta (Facebook, Instagram) and LinkedIn. Telegram works as a channel. Other systems connect through MCP servers.
Can an agent run on its own?
Yes. Save a task as an agent and start it on a schedule, from a webhook or when a new Gmail message arrives. You can watch each run's steps and results, and every run keeps logs and traces.
Build your first agent on a real task.
Book a demo and bring a job your team repeats. We will build it with you in Xagent.