Fibery MCP server should work in any MCP Client. Here are the most popular clients:
Our goal is simple: anything you can do in the Fibery UI, your AI assistant should be able to do through MCP.
Find and analyze. Query any database with filters and sorting, search across the whole workspace, see who changed what and when, and get the entities any existing view shows, with its filters applied.
"What changed in our workspace this week, and who changed it?"
Keep work moving. Create, update and delete entities, move them through workflow states, assign people, link related records, and add comments and mentions.
"Move everything I finished yesterday to Done and add a short comment on each."
Write and review documents. Read and edit descriptions and documents, add inline comments, or propose edits as suggestions you accept or reject. See Suggestions mode.
"Review this spec and propose your edits as suggestions."
Shape your workspace. Create spaces, databases and fields, including relations and formulas you describe in plain English. Set up workflows and build views: tables, boards, timelines, calendars, forms and more.
"Set up a CRM with Companies, Contacts and Deals, and a board of deals by stage."
Build reports. Create reports with charts, metrics and tables.
"Create a report with a chart of tasks closed per week."
Automate. Create and change automation rules and buttons by describing what they should do.
"When a task is marked urgent, assign it to the team lead."
Draw on whiteboards. Add shapes, sticky notes, text and sections.
"Draw our onboarding flow on a whiteboard."
Build custom apps. Assistants that work with files, such as Claude Code and Codex, can create and update custom apps. See Build Apps via MCP.
Work with files. Attach files from a link, or upload files from your computer in clients that can run commands on it.
"Attach this screenshot to the task I just created."
Import data. Get a link to the import wizard for CSV, Notion, Jira, Trello and other sources.
"Help me import our contacts from a CSV file."
Do multi-step work in one go. The assistant can write a short script that chains several steps, for example to loop over entities or total numbers across databases.
"For every project, count open tasks by assignee."