Memanto dashboard tutorial: memory and tool connections

This tutorial walks through Memanto's visual dashboard for cross-session AI agent memory. The presenter opens a local server and explains how the interface shares memory operations with the CLI. The overview shows the active agent, session status and backend health. Cloud and on-premises configurations have different settings; the speaker says the on-premises backend does not require an API key.

Most pages require an active agent and session. Each agent has a separate memory namespace, and activating one selects the target for dashboard, CLI and connected-tool operations. In the Playground, Answer generates a response from stored memories and displays its source. Recall performs semantic search without an LLM. The tutorial also covers document uploads and extracting memory candidates from conversations for review before saving.

Explorer lets users inspect stored memories and historical snapshots. History flags days with new memories but no conflict scan, while Conflicts compares contradictory or duplicate entries for resolution. Connections configures shared memory for coding tools through a selected project directory.

Migration accepts exported JSON or a provider API key and includes a dry-run preview. Imports write to the active agent, so the presenter stresses selecting the correct one first. The walkthrough ends with answer-model settings, retrieval controls, summary schedules and memory analytics.