v1.7.0
LatestFebruary 7, 2026Jules AI, MCP Server & Text Overhaul
A massive update packed with new tools. Jules AI is an early alpha experiment that can generate poster layouts from a text prompt — designs aren't great yet, but it's a fun starting point to iterate from. Plus: MCP server for AI agents, a complete docs site, interactive text tools, auto-layout groups, and a mountain of polish.

- Jules AI (early alpha) - describe a poster and Jules takes a shot at building it. Designs still need work, but it's a fun starting point to riff onfeature
- MCP server - any AI agent (Claude Code, Cursor, etc.) can create Efecto posters programmaticallyfeature
- Interactive text tool - click to create point text, drag to draw text frames (like Figma)feature
- Flex groups with auto-layout - group layers with flexbox direction, gap, padding, and alignmentfeature
- Full documentation site at efecto.app/docs with interactive API explorerfeature
- Animated text layers - 12 animation types including wave, typewriter, glitch, neon, and morefeature
- Smart model selection - Sonnet 4.5 for first drafts, Haiku 4.5 for fast iterationsfeature
- Jules uses flexbox layouts, groups, and design recipes to structure postersfeature
- Live streaming transcript so you can watch Jules think and build in real timefeature
- One-command MCP install - `npx efecto-mcp` sets everything up automaticallyfeature
- MCP text animations - all 12 animation types available via MCP toolsfeature
- MCP group support - create and manage grouped layers with auto-layout via AIfeature
- Docs search with Cmd+K and copy-page button on every pagefeature
- Complete effect documentation - every ASCII, dither, halftone, glitch, and art effect with parametersfeature
- Public API - state encode/decode, validation, async image rendering, and JSON schema endpointsfeature
- Text frame model - area text with auto-wrap and vertical alignment, separate from point textfeature
- Centralized typography engine - single source of truth for text measurement across canvas, renderer, and APIfeature
- Figma-style layer arrangement - drag to reorder, group/ungroup, and local-axis resize handlesfeature
- Duotone color presets for mono halftone - choose from curated ink + paper combinationsfeature
- Visual shape selector for halftone dots - circle, square, diamond, and line shapesfeature
- 3:4 aspect ratio - new portrait format option across canvas, URL sharing, and APIfeature
- 23+ new video samples added to the media libraryfeature
- Variable font support with proper Google Fonts loadingfeature
- Cmd+. keyboard shortcut to toggle sidebarsfeature
- User-friendly sliders - all effect and post-process controls now show human-readable values (degrees, percentages, whole numbers)improvement
- Editable slider values - click any slider value to type an exact numberimprovement
- NumberInput with stepper - shift+arrow for fine control, tab between fieldsimprovement
- Media browser performance - virtualized grids, image caching, skeleton loading, and static video thumbnailsimprovement
- Halftone tab moved before glitch in the effect panel for better discoverabilityimprovement
- Dither palette navigation with arrow keys and increased max point size to 20improvement
- Glitch art blog post with tips and techniques at efecto.app/blogimprovement
- MCP landing page at efecto.app/mcp with install instructions and docs linksimprovement
- Animated text now handles line breaks, emojis, spaces, and special characters correctlyfix
- Font loading race conditions resolved - fonts load reliably with proper fallbacksfix
- URL encoding fixes - background media, text defaults, and post-process settings all persist correctlyfix
- Better glitch defaults - Digital block opacity at 0, Weird chance at 66%, Chromatic shape squarefix
- Panel scrollbars hidden while keeping scroll functionalityfix
- Video play/pause race conditions and media reset on mouse move resolvedfix
- Comprehensive parameter validation and clamping across MCP server and APIfix









