An AI agent framework designed for teams, communities, and multi-user environments, Spacebot addresses the limitations of single-session AI agents by enabling concurrent, non-blocking thinking, execution, and response. It targets users needing robust, scalable AI agents capable of handling multiple users and tasks simultaneously, offering a significant benefit in responsiveness and throughput over traditional architectures.

How It Works

Spacebot's core innovation lies in its decomposition of a monolithic AI agent into five specialized, concurrently running processes: Channel (user interaction), Branch (for thinking/planning), Worker (for task execution), Compactor (context management), and Cortex (system oversight). This architectural choice allows the agent to perform multiple operations—like responding to one user, executing a task for another, and thinking about a third—simultaneously without blocking. This delegation model is key to its ability to scale to hundreds of users and concurrent conversations.

Quick Start & Requirements

• Primary install/run: Self-hosted via a single Rust binary (cargo build --release). One-click deploy available via spacebot.sh. A Docker image is also provided.
• Prerequisites: Rust 1.85+ (via rustup), an LLM API key from a supported provider (e.g., Anthropic, OpenAI, OpenRouter), or Anthropic OAuth credentials.
• Setup: Requires cloning the repository, building the binary, and configuring config.toml. No server dependencies or microservices are needed for self-hosting.
• Links: Repository: https://github.com/spacedriveapp/spacebot. General documentation links mentioned include How It Works, Architecture, Quick Start, Tech Stack, and Docs.

Highlighted Details

• Advanced Memory System: Employs a typed, graph-connected knowledge system (Fact, Preference, Decision, etc.) with hybrid recall (vector similarity + full-text search via Reciprocal Rank Fusion) and direct file import.
• True Concurrency: Manages simultaneous interactions across multiple users and channels via its process-based architecture, featuring message coalescing for rapid-fire inputs.
• Extensible Task Execution: Includes built-in workers for shell commands, file operations, program execution, and headless browser automation (Chromiumoxide). Supports an extensible skill system via skills.sh.
• Multi-Agent Support: Allows running multiple distinct agents with isolated identities, memories, and permissions from a single binary.
• Sophisticated Model Routing: Features a four-level routing system to select optimal LLMs for different tasks, including prompt-level scoring, fallbacks, and support for custom providers.

Maintenance & Community

Contributions are welcomed, with guides provided for code style and implementation. Specific community channels (e.g., Discord, Slack) or notable contributors are not detailed in the README.

Licensing & Compatibility

The project is licensed under the FSL-1.1-ALv2 (Functional Source License), which converts to Apache 2.0 after two years. This license has specific terms that users should review for compatibility, particularly concerning commercial use or integration into closed-source projects.

Limitations & Caveats

The primary caveat for adoption is the FSL license, which has a conversion clause to Apache 2.0 after two years. The README emphasizes Rust's safety features for infrastructure, but does not explicitly list known bugs, alpha status, or unsupported platforms.