DeepClaude offers a high-performance, self-hostable API and chat UI that merges DeepSeek R1's Chain of Thought (CoT) reasoning with Anthropic Claude's creative and code generation abilities. It targets developers and power users seeking to leverage the distinct strengths of both LLMs through a unified, private, and configurable interface, enabling advanced reasoning alongside robust coding and conversational tasks.

How It Works

DeepClaude employs a Rust-based backend for low-latency inference, orchestrating calls to both DeepSeek R1 and Anthropic Claude models. It streams responses from both models in a single, unified output. This approach aims to combine DeepSeek R1's advanced self-correction and reasoning capabilities with Claude's superior code generation and creativity, providing a synergistic effect that neither model achieves alone. The architecture prioritizes privacy by using the user's own API keys (BYOK) without storing or logging sensitive data.

Quick Start & Requirements

• Install: Clone the repository (git clone https://github.com/getasterisk/deepclaude.git) and build with cargo build --release.
• Prerequisites: Rust 1.75+, DeepSeek API key, Anthropic API key.
• Configuration: Create a config.toml file with server and pricing settings.
• API Usage: See API Docs for examples.

Highlighted Details

• Zero latency streaming responses from combined models.
• Private and secure BYOK architecture with local API key management.
• Highly configurable API and interface.
• Combines DeepSeek R1's reasoning with Claude's creativity and code generation.

Maintenance & Community

The project is developed by Asterisk. Further community engagement details (Discord/Slack, roadmap) are not explicitly provided in the README.

Licensing & Compatibility

Licensed under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The project requires specific API keys for both DeepSeek and Anthropic models to function. While the API is described as "Stable," the project's self-hosting nature implies a reliance on user-managed infrastructure and potential setup complexity.