openviking-memory

Long-term memory for AI agents is added by automatically capturing conversational facts and recalling relevant context before the agent responds, improving context awareness without manual intervention.

Agent memory recall works by automatically searching for and injecting relevant past memories into the context before the AI agent responds, utilizing configurable recall limits and score thresholds.

Yes, you can manually store or delete specific facts in AI agent memory, allowing explicit control over what information is retained alongside the automatic capture features.

Semantic memory capture extracts memories based on overall meaning, while keyword-based capture targets specific terms, both automatically extracting important information from user messages to build context.

Configuring recall limits and score thresholds for agent memory allows you to control the volume and relevance of past context injected before responses, ensuring personalized and accurate agent memory recall.

Yes, this AI memory solution operates without external dependencies, providing automatic conversational fact capture and recall for agents independently within the existing environment.