Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
FIGMA_PERSONAL_ACCESS_TOKENYesPersonal Access Token for Figma REST API to resolve published component keys. Generate in Figma desktop: Profile picture → Settings → Personal access tokens.

Capabilities

Features and capabilities supported by this server

CapabilityDetails
tools
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
mimic_statusA

START EVERY SESSION HERE. Returns Figma plugin connection status, the current build phase (0=idle through 5=report), enforcement profile, cached DS counts, knowledge store summary (learned components/patterns/gaps/rules), and a contextual hint telling you exactly what to call next. Also clears build-interrupt state after a plugin reconnect and warns if a completed build is missing its mandatory report. No params — call with no arguments.

mimic_discover_dsA

Complete DS discovery in two steps. Step 1: call with fileKey — discovers variables, text styles, components via plugin API, caches everything, stays at Phase 1. Step 2: call again with communitySearchResults (library names from Figma MCP search_design_system) — verifies no community libraries were missed, then advances to Phase 2 (build-ready). Build tools are BLOCKED until Step 2 completes. If a community library's variables are not discoverable via the plugin API (communityVariablesRequired response), fetch them via Figma MCP search_design_system and pass as externalVariables.

mimic_ds_assets

Manual escape hatch for DS discovery/preload/finalize — the low-level building blocks mimic_discover_ds already orchestrates automatically. Use this ONLY when the automatic flow needs a targeted redo: re-discovering one asset kind after a partial load, preloading variables/styles/fill-styles fetched some other way, or manually finalizing the enforcement profile. Params: action ("discover"|"preload"|"set_defaults"), kind ("styles"|"variables"|"components"|"fill_styles", required for discover/preload), plus action-specific params (fileKey for discover, styleKeys/variables/libraryFileKey for preload, dsMode for set_defaults). Workflow position: DS Discovery / Style Inventory (Phase 1-2), before building.

figma_list_dsA

Lists cached DS assets from the local cache — text styles, fill (color) styles, or variables. Use to browse what mimic_discover_ds already cached before binding: find a textStyleId for figma_create_text, a fillStyleId for figma_create_shape, or a variable path for figma_update_node. Params: kind ("text_styles"|"fill_styles"|"variables", required), filter (keyword, fill_styles only), category (text|background|border|foreground|spacing|radius, variables only). Read-only.

mimic_map_componentsA

Maps HTML element types (button, input, badge, table, tab, avatar, dropdown, textarea, header, footer, sidebar, ...) to DS component keys for the current build. Call once after mimic_discover_ds with all section-level + control element types. With FIGMA_TOKEN configured, one call is enough — all library components are pre-cached and missing types are confirmed gaps. Without a token, call again with librarySearchResults (from Figma MCP search_design_system) to close the loop. Workflow position: Phase 1-2, right after DS discovery, before figma_insert_component.

figma_create_frameA

Creates an auto-layout container frame (layout primitive) — the fallback when NO DS component exists for a section or layout. Every fill/padding/gap/radius binds to DS variables; component-like names are gated (pass confirmedNoComponent + primitiveOverrideReason for confirmed gaps). Name frames after their HTML role ("Header Section", "Card: Revenue"). Key params: parentId (omit for page-level artboard), direction, layoutSizing*, *Variable bindings, GRID mode. Check mimic_map_components FIRST. Phase 2+.

figma_create_textA

Creates a standalone text node bound to a DS text style (textStyleId — accepts style name or key) and DS color variable (fillVariable) — both are mandatory when the DS has them. Use for text OUTSIDE components; text inside a component instance uses figma_component_text instead. Strips hardcoded line breaks (container width controls wrapping). Name nodes after their HTML role ("Page Title", "Card: Revenue Label"). Params: parentId + content required. Phase 2+.

figma_create_shape

Creates a rectangle or ellipse primitive node bound to DS variables. Use for dividers, backgrounds, chart bars/grid-lines/data-points, avatars-as-circles, or any shape with no DS component. Params: shape ("rectangle"|"ellipse", required), parentId (required), name, width, height, fillStyleId/fillVariable/fill, and for rectangle: cornerRadius(Variable), stroke(Variable), strokeWeight, layoutPositioning, layoutSizing*; for ellipse: arcData ({startingAngle, endingAngle, innerRadius} — donut/pie segments). Requires Phase 2. Prefer a DS component over a shape when one exists.

figma_create_svgA

Creates a node from an SVG string in Figma. Useful for icons and custom graphics. Returns unboundChildren — a list of child nodes that need DS variable bindings. You MUST apply figma_update_node (op: "fill") to every unbound vector and figma_update_node (op: "text_style" then op: "fill") to every unbound text. Leaving unbound children breaks DS compliance and light/dark mode.

figma_insert_component

Imports and inserts a DS component instance into the document. Returns componentKey/nodeId, configurationChecklist (booleans to enable, text nodes to override, variants to set), and _autoApplied variant defaults replayed from the knowledge store when a confirmed recipe exists. Params: componentKey (required, from mimic_map_components/discovery), parentId (required), name (instance name override), importMode ("component"|"componentSet", usually inferred), applyRecipe (false to skip auto-replay). Workflow position: Phase 3 build, after mimic_map_components. Follow up with figma_component_text and figma_set_variant.

figma_component_textA

Overrides text on a component instance in one call — pass an overrides array (a single override is just a one-item array). Each override sets one text node via textNodeName (matches by name — use for most cases) or textNodeId (exact match — use when configurationHints.textNodes gave you IDs, e.g. components with repeated node names). Use after figma_insert_component; configurationHints.textNodes lists every node that needs overriding. Saves N-1 tool calls vs. one call per node. Requires Phase 2.

figma_set_variant

Sets variant properties (e.g. Size, State, Color) on a component instance. Use figma_inspect (target: "variants") first to see available properties/values for the component set. Params: nodeId (required), properties (required, key-value map e.g. { "Size": "Large" }). Requires Phase 2. Tracks replay success/failure for the knowledge store's recipe learning.

figma_swap_main_component

Swaps or replaces a component instance's underlying component. mode="swap" (default) preserves the instance's overrides (text, variants) — use when switching to a close variant of the same component. mode="replace" removes the instance and creates a fresh one at the same position — use when swapping to an unrelated component where old overrides would be invalid. Params: nodeId (required), newComponentKey (required), mode ("swap"|"replace", default "swap"). Requires Phase 2.

figma_manage_slotA

Fills or resets a SLOT-type component property on an existing instance (Figma Slots, GA June 2026). action="fill" inserts a DS component instance into the slot (use configurationHints.slotProperties from figma_insert_component to find slotName). action="reset" restores the slot's default content. Params: nodeId (required), slotName (required), action ("fill"|"reset", default "fill"), componentKey (required for action="fill"). Requires Phase 2. Slot fills are recorded for build-report visibility only — never auto-replayed.

figma_update_nodeA

Mutates an existing Figma node or the editor's UI focus, dispatched by op. Ops: "text" (set text content), "text_style" (apply DS text style), "fill" (fillStyleId/fillVariable/raw fill), "layout" (sizing, padding, gap, alignment), "visibility" (show/hide), "position" (absolute x/y), "restyle" (artboard/frame fill+radius+stroke+padding+gap in one call), "move" (reparent via parentId+index), "select" (select+scroll into view), "page" (switch page by name/id). All ops except "select"/"page" require Phase 2 (DS discovered) and validate any *Variable path against the cached DS before sending. Use figma_delete_node (separate, destructive tool) to remove a node instead.

figma_variable_modesA

Sets the variable mode (e.g. light/dark) on a node — required on every new artboard, or DS variables render as black. By default sets the mode on ALL variable collections at once (pass modeIndex only). Pass collectionName to target a single collection instead. Params: nodeId (required, typically the artboard), modeIndex (required, 0=default/light, 1=dark), collectionName (optional). Requires Phase 2.

figma_delete_nodeA

Deletes a node from the Figma document. NEVER deletes artboards (top-level frames) — only child nodes within an artboard. To rebuild a screen, create a NEW artboard alongside the existing one. Destructive and irreversible via the API — verify the nodeId with figma_inspect first if unsure.

figma_inspectA

Reads Figma document state — node properties, children, parent, text detail, pages, top-level page nodes, selection, component variants, or a build-manifest section lookup. Read-only, never blocked by phase gates. Use to verify state before/after edits, discover node IDs, or check what a component set supports. Params: target (required: "node"|"children"|"parent"|"text"|"pages"|"page"|"selection"|"variants"|"section") selects the read; nodeId (node/children/parent/text), depth (children, default 1), componentSetKey (variants), sectionName (section — HTML section name from the last build).

mimic_ai_knowledge_readA

Loads the knowledge store: learned component recipes, layout/pattern data, DS gaps, user-defined rules, and meta stats (build counts, replay savings). Call at the start of a session to see what Mimic has already learned about this DS. Params: format ("json" default — structured data; "design_md" — renders the current DS + recipes as a DESIGN.md-style markdown string for documentation/handoff).

mimic_ai_knowledge_writeA

Saves a component recipe, layout pattern, DS gap, or user-defined design rule to the persistent knowledge store — the mechanism that lets Mimic learn across builds. Use "rule" when the user corrects build behavior in a generalizable way (e.g. "cards always have a header + content frame"); other types are usually written automatically by the build pipeline. Params: type ("component"|"pattern"|"gap"|"rule", required), id (unique key, required), data (entry payload, required).

mimic_generate_build_report

Compiles the current build session into a structured report — DS component usage, primitives with justifications, DS changes/staleness, gap recommendations, binding-quality failures, efficiency (tool calls, cache hits, replay savings), and design-rule compliance. MANDATORY at the end of every build, before responding to the user — advances phase to 5 and resets per-build counters. No params.

mimic_compute_chart

Computes chart geometry (bar positions, SVG paths, arc angles, grid lines, axis labels) from raw data WITHOUT building anything in Figma — pure math, no bridge/Figma calls, usable at any phase. Use when you need coordinates/paths to hand-build a chart yourself; use mimic_build_chart instead when you just want the finished chart created in one call. Supports chartType: bar, horizontalBar, donut, line, radar, scatter, heatmap. Params: chartType (required), data (required, points), dimensions (chart-type-specific sizing).

figma_validate_ds_complianceA

Recursively checks a node (and its children) against the current DS enforcement profile — flags hardcoded colors/fonts instead of variables/styles, missing text style bindings, and other DS violations. Use during Phase 3/4 QA, especially at the 20-op build checkpoint, before generating the report. Params: nodeId (required, typically the artboard root).

mimic_pipeline_resolve

Classifies build input as a URL, local file path, or raw HTML string and returns the resolved content ready for building. Use FIRST when the user gives you a link, a file path, or ambiguous input instead of raw HTML — it tells you what you actually received. Params: input (string, required). Workflow position: before mimic_discover_ds, at the very start of a build.

mimic_build_tableA

Bulk table builder — creates an entire data table in ONE call: column frames, DS Table header cell + Table cell components, variants (cellVariants for per-value badge colors), text, and consistent row height. Use for ANY HTML data table instead of cell-by-cell insertion; reduces 200+ tool calls to 1. Requires table cell components in the DS (returns creation guidance if missing). Key params: parentId, columns (header/style/cellVariants), rows ("text|supporting" syntax), cellHeight. Phase 2+.

mimic_build_chartA

Bulk chart builder — creates an entire chart in ONE call: container, visualization (native rectangles or stroke-free SVG), axis labels, grid lines, and legend, all bound to DS variables and text styles. Use for ANY HTML chart instead of hand-building; reduces 30-50 tool calls to 1. chartType: bar, line, donut, radar. Key params: parentId, chartType, title, data, dimensions, colors (neutral DS data palette — never Brand/Success/Warning/Error). Use mimic_compute_chart for geometry-only. Phase 2+.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/miapre/mimic-ai'

If you have feedback or need assistance with the MCP directory API, please join our Discord server