Fibery MCP Server Tools References

Guide

Tools by Category

Explore Your Workspace

Tool	Description
`schema`	Lists every space and database in the workspace. The AI calls this first before doing almost anything, to know what exists.
`schema_detailed`	Zooms into a specific database and shows all its fields, their types, and their possible values (e.g. all workflow states). The AI needs this before reading or writing data accurately.
`get_me`	Returns the currently logged-in user's name, email, and ID.
`search_guide`	Searches the official Fibery documentation to answer questions about how Fibery itself works (not your data).
`get_fibery_skill`	Loads the full reference guide for a Fibery skill domain, including syntax, configuration shapes, and workflows for related tools. Available domains: `schema`, `data`, `query`, `documents`, `views`, `reports`, `automations`, `whiteboard`, and `custom-apps-dev`.

Search & Query

Tool	Description
`query`	The main read tool. Fetches entities from any database with filtering, sorting, pagination, and aggregation. Supports sub-queries (e.g. "give me each feature and its linked bugs").
`search`	Full-text keyword search across the entire workspace — entity names, document content, and comments. Returns a ranked list of matches.
`search_history`	Looks up the activity log: what was created, updated, or deleted, by whom, and when. Supports filtering by database, entity, author, date range, action type, and schema changes. By default it covers the last 24 hours and skips automatic changes (automations, integrations, formulas, auto-linking). Ask for a specific period or for automatic changes when you need them. One request can cover up to 12 months.
`query_views`	Lists saved views (boards, timelines, grids, documents, etc.) and their configuration.
`fetch_view_data`	Executes a saved view and returns the actual entities it would show, with the view's own filters applied.
`fetch_by_url`	Fetches entity or view data from a Fibery URL and renders it as markdown.
`execute_code`	Runs a short JavaScript script that calls the other tools as functions. The AI uses it for multi-step tasks — loops, aggregations, joining data across databases — in one call instead of many.

Permissions & Access

Tool	Description
`display_schema_capabilities`	Reports what the current user can and cannot do at the space and database level across the workspace.
`display_entity_capabilities_via_sharing`	Reports per-entity access derived from sharing: entities reached through direct sharing, access templates, and assignment rules based on People fields.

Create & Update Entities

Tool	Description
`create_entities`	Creates one or more new entities in a database (bugs, features, tasks, customer requests, etc.).
`update_entities`	Updates fields on existing entities. Only the fields you specify are changed.
`delete_entities`	Permanently deletes entities. Cannot be undone.
`set_state`	Moves an entity to a different workflow state (e.g. "In Progress" → "Done").
`get_entity_links`	Returns direct Fibery web links to entities by their public IDs (like #13961).
`add_collection_items`	Adds items to a collection field: assign a person, attach a tag, link a related entity.
`remove_collection_items`	Removes items from a collection field.
`add_comment`	Adds a comment to an entity or replies to an existing comment using `parentCommentId`. The parent comment must belong to the same entity.

Documents & Files

The editing tools change the document directly by default. insert_document_blocks, set_block_text, replace_block_text, replace_document_blocks, and delete_document_blocks also accept mode: "suggest" to record the whole batch as suggestions (tracked changes) a reviewer accepts or rejects; each suggested change can carry a rationale comment shown next to it.

Tool	Description
`read_document`	Reads a document as a flat list of addressable blocks with stable IDs, plus any inline comment threads and pending suggestions (tracked changes). Read it before editing existing content to obtain block IDs and understand the current structure. Inserting at the document root does not require an existing block ID.
`insert_document_blocks`	Inserts new blocks (headings, paragraphs, lists, tables, images, etc.) from markdown. Insert before/after an existing block or at the start/end of a container. To insert at the document's start or end, use `parent` with `blockId: null` and `position: "start"` or `"end"`.
`set_block_text`	Rewrites the inline content of one or more existing text blocks from markdown in a single call, preserving each block's type, attributes, ID, and position. Each replacement must contain a single block's inline content; use `insert_document_blocks` to add blocks or `replace_document_blocks` to rewrite a block as several.
`replace_document_blocks`	Replaces a block, with all its children, by new blocks from markdown in the same place: e.g. a paragraph into a heading plus a list. Also used to fill an empty document. Tables can be replaced only as a whole.
`replace_block_text`	Replaces one exact text occurrence inside a block, leaving the rest untouched. The preferred tool for small fixes — surrounding formatting and comments survive. Plain text matches regardless of formatting, and the replacement can carry inline markdown (bold, links, strikethrough).
`move_document_blocks`	Moves blocks (with all their children) to a new position in the document.
`delete_document_blocks`	Deletes blocks, each with all its children.
`set_block_attrs`	Changes block attributes: heading level, task done/todo state, code block language, callout icon, etc.
`add_inline_comments`	Adds inline comments to selected text — the same comment threads users see in the editor. Each comment must be anchored within one block; multiple comments can be added in a single call.
`reply_document_comment`	Adds replies to existing inline comment threads. Use each thread's root comment ID; multiple replies can be added in a single call.
`add_suggestion_comments`	Adds comments anchored to pending suggestions (tracked changes) — to question or discuss a suggestion instead of resolving it. The thread follows the suggestion until it is accepted or rejected, then lives on as an ordinary text comment.
`resolve_document_suggestions`	Accepts or rejects pending suggestions, specific ones by ID or all at once. Resolving changes the document directly: accepting a deletion removes the content, rejecting it keeps it; accepting an insertion keeps it, rejecting it removes it.
`get_entity_mention` / `get_user_mention`	Builds a "live" entity reference or user mention to embed into document content via the editing tools.
`add_file_from_url`	Downloads a file from a URL and attaches it to an entity's file field.
`prepare_file_upload`	Uploads a file from your computer to Fibery. Works only in clients that can run shell commands (Claude Code, Codex CLI, Cursor, etc.): the tool returns a `curl` command, and the file goes straight to Fibery, not through the chat. Then attach it with `attach_files` or embed it in a document.
`attach_files`	Attaches files already stored in Fibery to an entity's file field: a file uploaded with `prepare_file_upload`, a file from another entity, or a file attached in chat.
`download_file`	Fetches a file attachment by its secret and returns a signed download link valid for about an hour, plus file metadata. Small images and supported text files are also returned inline for the AI to inspect.
`get_files_meta`	Lists file attachments on entities.

Structure (Spaces, Databases, Fields)

Tool	Description
`create_space` / `delete_space`	Creates or removes a top-level space.
`create_databases` / `rename_databases` / `delete_databases`	Manages databases within a space. Renaming can also move a database to a different space.
`create_primitive_fields`	Adds simple fields: text, number, date, date range, checkbox, URL, email, phone, location, or rich-text document.
`create_relation_fields`	Creates a link between two databases (e.g. Bug ↔️ Customer Request), with a chosen cardinality (one-to-many, many-to-many, etc.).
`create_single_select_fields` / `update_single_select_fields`	Creates or modifies a single-choice dropdown field with colored options.
`create_multi_select_fields` / `update_multi_select_fields`	Creates or modifies a multi-choice tag-style field.
`create_workflow_field` / `update_workflow_field` / `delete_workflow_field`	Manages the workflow (state) field: defines stages and their types (Not started / Started / Finished).
`create_formula_field` / `update_formula_field`	Creates or updates a calculated field. You describe what it should compute in plain English; Fibery generates the formula.
`rename_fields` / `delete_fields`	Renames or removes fields from a database.
`create_comments_fields` / `delete_comments_fields`	Enables or disables comments on entities in a database.
`create_avatars_fields` / `delete_avatars_fields`	Enables or disables avatar images on entities.
`create_icon_fields` / `delete_icon_fields`	Enables or disables emoji icons on entities.
`create_files_fields`	Adds a file attachment field to a database.

Automations

Load get_fibery_skill({skill: "automations"}) first for trigger, filter, and action formats. Automation rules run on triggers; buttons run when clicked. Automation tools are not available in ChatGPT and Codex.

Tool	Description
`query_automations`	Lists the automation rules or buttons for a database, including triggers, filters, actions, and IDs needed for updates.
`get_automation`	Fetches one automation rule or button by ID, including its triggers, filters, and actions.
`display_automations_schema`	Lists the available actions and their arguments for a database and automation type. Call it before creating an automation or adding/changing its actions.
`create_automation`	Creates an automation rule or button for a database. Rules require at least one trigger; buttons have no triggers or filter.
`update_automation`	Modifies an existing rule or button: rename, enable/disable, replace triggers or filters, and add, change, or delete actions. Fetch the automation and its action IDs first.
`delete_automation`	Permanently deletes an automation rule or button. Use `update_automation` to modify one.
`get_connected_app_accounts`	Lists connected external app accounts, such as Slack or email. Use an account's ID as `appAccountId` when an automation action requires an external app account.

Views

Tool	Description
`create_view`	Creates a new view: grid, list, board, timeline, calendar, gantt, gallery, map, feed, embed, form, standalone document, or whiteboard (`canvas`). Documents and whiteboards are created empty; add their content with the document block tools or `execute_whiteboard_commands`.
`update_view`	Edits an existing view's name, description, space, or configuration. A supplied `config` replaces the previous configuration entirely. For document views, it only renames, moves, or edits the description; use document block tools to edit the body.
`delete_views`	Removes views (entity data is not affected).

Whiteboards

Create a new empty whiteboard with create_view using viewType: "canvas", or find an existing one with query_views. Load get_fibery_skill({skill: "whiteboard"}) before editing its elements. In ChatGPT and Codex, whiteboards can be created and found but not edited: execute_whiteboard_commands is not available there.

Tool	Description
`execute_whiteboard_commands`	Creates, updates, or deletes whiteboard elements, including shapes, text, links, sticky notes, sections, and freehand strokes. Returns the whiteboard URL and an `idMap` mapping placeholder element IDs to real IDs for subsequent edits.

Reports

Tool	Description
`get_reports_list`	Lists all report views in the workspace.
`get_report`	Fetches a report by ID with its tabs, sources, and configuration.
`create_report`	Creates a report with sources and a title.
`update_report`	Updates a report's title, sources, or both.
`display_report_schema`	Shows the report source schema for one or more databases.
`add_chart_tab` / `add_metric_tab` / `add_table_tab`	Adds a chart, metric, or table tab to a report.
`update_tab`	Updates a tab's title, filters, and, for chart tabs, chart type and palette.
`remove_tab`	Removes a tab from a report.
`update_dimension` / `remove_dimension`	Modifies or removes a dimension within a tab.

Custom Apps

Tool	Description
`get_custom_apps_list`	Lists the custom apps in the workspace (small React apps embedded in Fibery views).
`create_custom_app`	Creates a new custom app scaffolded from the starter template, ready for development.
`create_custom_app_dev_token`	Issues a short-lived access token for developing a custom app locally.

Import

Tool	Description
`get_connectors_list`	Lists all available built-in integrations (Trello, Notion, Jira, CSV, etc.).
`get_manual_import_link`	Generates a URL to the Fibery import wizard for a specific connector and target database.

Not available in ChatGPT and Codex

When you connect from ChatGPT or Codex (web, CLI, or IDE), the server hides these tools:

Automations: query_automations, get_automation, display_automations_schema, create_automation, update_automation, delete_automation, get_connected_app_accounts
Whiteboards: execute_whiteboard_commands
Code execution: execute_code

Why: OpenAI doesn't approve tools that discover their available actions dynamically, and the automation and whiteboard tools work exactly that way. We plan to rework them so OpenAI can approve them. Until then, they are unavailable in these clients.

To view or change automations or edit whiteboard content, use the Fibery app directly or Fibery AI. Every other tool on this page works in all MCP clients.

Use Cases

Each use case shows a natural prompt, the main tools the AI calls behind the scenes, and what comes back. Schema discovery and loading the relevant guide with get_fibery_skill may also be needed.

Find Things

Prompt	Tools	Result
"Show me all urgent bugs that aren't done yet"	`schema_detailed` → `query`	Table of urgent bugs with name, state, and assignee — excluding Done, Tested, Won't Fix
"What did the team work on this week?"	`search_history` with the requested week	Activity for the requested week, grouped by database with links. Automatic changes are excluded unless explicitly included.
"Find everything related to inline comments"	`search`	Ranked list of matching bugs, features, customer requests, and documents across the whole workspace
"Give me a full picture of the history migration feature"	`query` → `read_document` → `query`	Feature state, owner, planned dates, linked open bugs, and the full description document
"Who has the most open bugs assigned to them?"	`schema_detailed` → `query`	Team members ranked by open bug count, excluding Done/Won't Fix/Tested
"Which Icebox features have the highest customer demand?"	`schema_detailed` → `query` → `get_entity_links`	Features sorted by Highlights Score descending, with direct links

Create & Update Things

Prompt	Tools	Result
"Create a bug: form fields lose their value when switching tabs. Assign it to Nastya and mark urgent."	`schema_detailed` → `create_entities` → `query` → `add_collection_items` → `update_entities`	Bug created, Nastya assigned, urgent set. Confirmation with a link.
"Move all Implemented features to Done"	`schema_detailed` → `query` → `set_state`	All matching features transitioned. Summary: "Moved 8 features to Done: Handle Large Datasets…"
"Write a structured description for the inline comment icons bug"	`search` → `query` → `read_document` → `insert_document_blocks`	Description written with Summary, Steps to Reproduce, Expected vs. Actual, Environment sections
"Add a Severity field to bugs — Critical, High, Medium, Low"	`schema` → `create_single_select_fields`	Severity field created with four color-coded options, Medium as default
"Link this customer request to the Figma plugin feature"	`search` → `query` → `add_collection_items`	Request linked to feature; both entities now reference each other
"Set up a new Design space with task tracking linked to features"	`create_space` → `create_databases` → `create_workflow_field` → `create_primitive_fields` → `create_relation_fields` → `create_comments_fields`	Space created with two databases, workflow, Figma Link field, relation to features, and comments enabled

Multi-Step Workflows

Prompt	Tools	Result
"Prepare my standup — what did I do yesterday and what's stuck?"	`get_me` → `search_history` → `query`	Yesterday's activity, current in-progress items, and items with no update in 3+ days flagged as at risk
"Audit features with no owner and no planned dates"	`schema_detailed` → `query` → `get_entity_links`	List of features missing owner or dates, grouped by gap type, with direct links
"Write this week's release notes from what shipped"	`search_history` with the requested week → `search` → `read_document` → `insert_document_blocks`	New dated section appended to the Release Notes document with shipped bugs and features found in the requested period. Include automatic changes when workflow transitions are automated.
"Triage new customer requests — match each to the right feature"	`query` → `read_document` → `query` → `add_collection_items` → `set_state`	Requests matched to features and moved to In Review; unmatched ones flagged for manual review
"Give me a sprint health report — what's at risk?"	`query_views` → `fetch_view_data` → `query` → `search_history` → `get_entity_links`	Total items by state, items with no recent activity flagged as at risk, overdue items highlighted

Automations & Whiteboards

Prompt	Tools	Result
"Show me the automation rules for Bugs"	`get_fibery_skill` → `query_automations`	Rules listed with their triggers, filters, and actions
"When a bug is created, add a comment asking the assignee to triage it"	`get_fibery_skill` → `display_automations_schema` → `create_automation`	A rule that adds the triage comment when a bug is created
"Create a sprint retrospective whiteboard with Went well, Improve, and Actions sections"	`create_view` → `get_fibery_skill` → `execute_whiteboard_commands`	A whiteboard with the three sections and a direct link