Remembra is an AI memory control plane for agent workflows. It provides a shared memory bank for Claude, Cursor, Codex, OpenAI agents, and automations, with provenance, temporal context, entity resolution, and a dashboard for operating the memory layer. The page describes it as a shared brain and as an operating layer between tools and agents, aimed at work where context is scattered across multiple agents and interruptions.
Its features include memory extraction, temporal truth, graph recall, conflict review, and security controls. The system turns conversations into durable facts, entities, and deduped memories, with fact extraction, importance scoring, entity extraction, speaker attribution, and streaming progress for larger sessions. It also supports conflict review, temporal TTL presets, natural decay, as-of historical queries, stale-context cleanup, and a recall pipeline that combines semantic search, BM25, graph traversal, reranking, and token-aware context shaping. The page also mentions provenance, temporal context, relationships, and guardrails around team, key, and audit visibility.
Remembra is built for teams and organizations using AI agent systems and multi-tool workflows that need continuity, accountability, source-aware recall, and context that improves over time. It mentions meeting notes, code decisions, campaign context, launch plans, documents, CRM, Slack, issues, and agent handoffs as material that can live in the memory layer. The dashboard is described as including a visual memory browser, entity graphs, query debugging, teams, projects, keys, billing, activity visibility, and recall debugging. The API surface includes conversation ingest, recall debug, webhooks, inbox, meetings, and transfer endpoints, and the docs cover install guides, MCP setup, API reference, diagnostics, and integration paths.
Delivery is self-hosted, with a documented Docker self-host setup and MCP native tools for agent workflows. The page also references an MCP server and a one-command installer for supported tools. Pricing is shown as free on the landing page. A changelog, dashboard, docs, and GitHub presence are listed in the site navigation.
Remembra sits in PulseGate's Memory & skills category. It focuses on providing a unified, self-hosted memory layer for AI agents to share context and resolve entities across tools. It is built as an open-source project for AI developers and automation engineers. Remembra is open source under the MIT license. It ships for the command line and API, and it can be self-hosted.
Behind Remembra is Remembra, and it first shipped in 2026. The project is developed in the open on GitHub with 14 stars and 161 commits in the last 90 days. Key capabilities include memory control plane, entity resolution, and context sharing. It exposes integrations via an MCP server and a public API.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do