Waku Unifies AI Coding Agents in GPU-Accelerated Desktop Hub
Built with Rust and GPUI, the new local-first application consolidates fragmented AI agent CLIs into a single native interface.
Waku (v0.1.0) has launched as a native desktop application designed to provide a unified interface for the expanding ecosystem of local coding agent command-line interfaces (CLIs). By consolidating multiple AI agents into one environment, the tool eliminates the fragmentation developers face when switching between disparate provider tools.
Built with Rust and the GPUI framework—the same GPU-accelerated technology powering the Zed editor—Waku avoids the resource overhead typical of Electron-based wrappers. The application currently supports a wide array of agent CLIs, including Claude Code, Amp, Codex CLI, Cursor CLI, Grok Build, OpenCode, and Pi. Its architecture utilizes a native desktop RPC client paired with a standalone "waku-daemon" process, which manages Git operations, filesystem access, and SQLite data storage.
Solving Agent Fragmentation
As AI coding agents proliferate, developers have increasingly relied on a patchwork of CLIs, each with its own communication protocol. Waku addresses this by acting as a provider-neutral GUI that normalizes various protocols, such as JSON-RPC and stream-json, into a single, cohesive timeline. This allows users to manage multiple agent sessions, projects, and transcripts without requiring a cloud account.
One of the tool's most distinct technical features is the implementation of "conversation-aware checkpoints." By utilizing hidden git refs, Waku can simultaneously roll back both the state of the codebase and the provider's conversation history. This ensures that if an AI agent introduces a breaking change, the developer can revert the project and the AI's context to a previous known-good state in one action.
The Shift to Local-First Tooling
The release of Waku signals a broader trend toward "local-first" AI development tooling. By keeping all session data and transcripts on the user's local machine, the application prioritizes privacy and performance over cloud-dependent architectures. The choice of GPUI emphasizes this commitment to speed, targeting the high-performance standards of modern native editors to provide a more responsive user experience than traditional web-tech wrappers.
Open Standards and Future Outlook
Released under the GNU General Public License v3.0, Waku is positioned as an open-source utility for the developer community. As more AI providers release specialized CLIs, the project's ability to normalize these interfaces will be critical to its adoption. Future observers will likely watch for the addition of more agent providers and the evolution of its daemon-based architecture as it handles increasingly complex local project structures.