codependentai/resonant

Resonant combines identity persistence, relational memory, and MCP protocol integration into an agent framework—novel in combination but built entirely on top of Anthropic's Claude Agent SDK and MCP standard. At 23 stars with 19 days age and zero recent velocity, this is an early-stage project with minimal adoption signal. The core dependency on Claude Agent SDK creates direct platform risk: Anthropic is actively building multi-agent orchestration capabilities (Agents API, Tool Use, MCP expansion) and could absorb identity-persistence and relationship-tracking as native features within the SDK itself within 6-12 months. The relational AI / memory persistence space is also attracting VC-backed competitors (e.g., memory-centric agent frameworks, context managers) and established players (OpenAI with memory APIs, LangChain ecosystem enhancements). No evidence of production deployments, enterprise traction, or community contributors beyond the original authors. The framework is composable as a library but offers no switching costs—users are locked to Claude SDK regardless, making migration to a competing solution trivial if Anthropic adds native support. Defensibility is weakened by: (1) high dependency on a single platform (Anthropic), (2) lack of differentiating moat (identity + memory + MCP are individually commoditizing), (3) no data or network effects, (4) nascent community. Risk horizon is 1-2 years because Anthropic's roadmap visibility and rapid multi-agent capability expansion create moderate urgency for platform absorption. Not immediately threatened (6 months) due to low visibility, but vulnerability grows as the company scales the Agent SDK ecosystem.