EvoForge

About EvoForge

EvoForge is a local-first personal AI system designed to operate entirely on the user’s own computer. It is built on an original agent kernel written in Rust rather than wrapping external command-line interfaces, enabling deeper integration and control across all layers of operation. The system supports frontier-level coding and can see, hear, read, and generate content across multiple modalities including text, images, audio, video, and documents. It includes 98 native host tools for tasks such as reading documents, driving browsers, managing persistent shells, and rewriting code structurally. Context management is handled through compact budgets and tiered attachments to prevent model overflow. Coordination can be handled either through a lead model that assigns tasks or a swarm model where agents claim tasks from a shared board with leases and peer notes. Changes are tested in isolated worktrees with regression tests required to pass before integration, ensuring the codebase only improves over time. The director’s console allows users to direct the creation of finished films by managing narration, shot generation, and multi-track timelines with human approval at each stage.

Key features

• ✓Frontier-level coding with original kernel
• ✓Native support for text, image, audio, video, and document processing
• ✓Lead and swarm agent coordination models
• ✓Isolated worktree testing with regression gates
• ✓Multi-track director's console for film production
• ✓54 built-in original skills with measured performance
• ✓Persistent shell and browser automation tools
• ✓Self-evolving skill system with promotion gates

Use cases

• ✓Generating and editing multimedia content locally
• ✓Automating complex coding and system design tasks
• ✓Creating narrated films with human-in-the-loop direction

Frequently asked questions about EvoForge