Agent- and model-agnostic memory backend with recursive compression and LLM-powered distillation. Your agents never forget what matters.
TotalRecall provides persistent, structured memory for autonomous agents. It compresses conversation history through recursive summarization powered by large language models, retaining critical information while discarding noise.
The system is designed to be agent-agnostic and model-agnostic - it works with any AI framework and any LLM provider. Memory is organized in hierarchical layers that balance recall fidelity against token efficiency.