COEY builds intelligent content systems through GroundSlate, a local AI super app for macOS that turns AI assistants like Claude and ChatGPT into video editors. GroundSlate indexes media on the Mac, enabling AI-driven search, cutting, editing, and rendering at hardware speed with non-destructive MCP server commands. The company also offers professional services for workflow architecture, content production, and operated channels.
Funding
Funding not disclosed
Founders
Product
Problem
Content creators and media teams face a bottleneck when producing video: searching through raw footage, cutting precise moments, and rendering finished edits requires specialized software, manual effort, and often cloud-based AI processing that separates users from their local media. This slows production workflows and limits the ability to scale content output while maintaining creative control.
Solution
GroundSlate by COEY is a local AI super app for macOS that functions as an MCP server, enabling AI assistants like Claude, ChatGPT, and Codex to act as video editors with full context of the user's local media. The software indexes every word, face, and place in the footage on the Mac, allowing AI agents to search for specific moments, cut and assemble edits, duck background music, build titles, and render finished files at hardware speed. By keeping all processing local and providing 100% coverage of non-destructive commands, GroundSlate enables continuous 24/7/365 editing and rendering without cloud dependency. The system integrates with any MCP-compatible client, making it a flexible addition to existing AI-driven workflows.
Target Audience
Primary users are video content creators, media teams, B2C and B2B brands, agencies, and media companies needing to automate video production at scale on macOS, along with executive leaders seeking to implement automated content systems.
Features
- MCP server architecture with 100% non-destructive command coverage, allowing AI clients (Claude, ChatGPT, Codex) to control the editing process
- Local media indexing that identifies spoken words, on-screen text, faces, and places, returning exact time windows instead of whole clips
- AI-driven video assembly, including auto-ducking music under voice, title card generation, and rendering through the Mac's hardware (optimized for Apple silicon)
- Library management for thousands of clips and moments, with search functionality and activity tracking (e.g., 3,140 clips, 26,904 moments indexed)
- A suite of 412 built-in tools for tasks like clip resolution, quick-cut planning, overlay authoring, and concatenation rendering, all executed on the local machine
- Runs with any MCP client, ensuring interoperability with existing AI ecosystems