This open-source template provides a foundation for building a durable personal AI agent. It enables users to interact with their agent across multiple platforms including web chat, Slack, iMessage, and Linear, offering persistent chat threads and user-approved long-term memory saves. The project aims to deliver a unified and integrated AI assistant experience for power users and developers.

How It Works

The architecture comprises a Nuxt.js frontend and API layer (UI, Better Auth, SQLite database) and a backend agent runtime called Eve. Eve manages agent logic, tools, and skills, communicating with the Nuxt API for session state and memory. Vercel Connect facilitates integrations with services like Slack and Linear. This design allows for a single codebase to power diverse interaction surfaces, with a focus on durable sessions and explicit user control over memory storage.

Quick Start & Requirements

• Install/Run: Clone the repository, install dependencies with pnpm install, copy environment variables (cp .env.example .env), migrate the database (pnpm db:migrate), and start the development server (pnpm dev). Access the UI at http://localhost:3000.
• Prerequisites: Node.js 24+, pnpm.
• Environment Variables: Requires BETTER_AUTH_SECRET, BETTER_AUTH_URL, and INTERNAL_API_SECRET.
• Local Database: A fresh local database can be initialized with rm -rf .data/db && pnpm db:migrate.
• Links: Official GitHub Repository

Highlighted Details

• Multi-Platform Integration: Supports web chat, Slack DMs/mentions, iMessage (via Sendblue), and Linear issue tracking.
• Durable Sessions & Memory: Chat threads persist across sessions, and save_memory proposals require explicit user approval before storage.
• Long-Term Memory: Allows importing memory from other assistants (e.g., ChatGPT), with options to edit, delete, or have the agent propose updates.
• Skills: Includes a "Daily Summary" skill for morning briefings on active focus, assigned issues, and suggested actions.

Maintenance & Community

No specific details regarding notable contributors, sponsorships, or community channels (like Discord/Slack) are provided in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: The MIT license is permissive, generally allowing for commercial use and integration within closed-source projects.

Limitations & Caveats

As a template, the project requires significant customization to define a specific agent persona, integrate desired AI models, and configure external service connections. Specific limitations, unsupported platforms, or known bugs are not detailed in the provided README.