Questions, answered.

Institutional engineering memory for humans and AI agents. Your code remembers what changed; Recallium remembers why: the decisions, fixes, constraints and rules behind your software, captured while the work is done and recalled by every engineer and agent on the team. Connect your repositories to Recallium Cloud through the Model Context Protocol (MCP), and Claude Code, Codex, Cursor and Copilot all start from what the team already knows instead of re-learning the codebase each session.

Any client that speaks MCP: Claude Code, Claude Desktop, Cursor, Codex, VS Code, Windsurf and the JetBrains IDEs among more than 60 tools. If your agent harness can add an MCP server, it works with Recallium.

Mem0 remembers users. Recallium remembers engineering. Mem0 and Supermemory are memory APIs you build into an AI product so it remembers its users: both list unlimited end users on their plans, and both also ship Claude Code and Cursor plugins. Recallium is institutional memory for the team building the product: it keeps the decision, the alternatives it beat, the constraint and the fix, knows which decision is current, follows the work from design through debugging and shipping, loads team rules into every session, and links commits to the memories behind them, with teams and role-based access on Recallium Cloud. It is priced per engineer.

On LongMemEval-S, Recallium finds the right memory for 499 of 500 questions (99.8% hit@10) and returns 98.6% of all relevant sessions, from 3.8 results per question on average, abstention questions included. The full protocol, readers, judge, depth and run dates are published.

Yes. Recallium was called MiniMe MCP from August 2025 until it was renamed Recallium. Guides and articles that use the MiniMe name describe an older version.

Yes. Cloud Free is hosted, includes 500 memories a month until December 2026 for personal use, and needs no infrastructure to run. Upgrade to Pro for shared team memory.

Both are hosted on Recallium Cloud; join the waitlist, or use a pilot code if you have one. Cloud Free includes 500 memories a month until December 2026 for personal use; Cloud Pro adds shared team memory and raises that to 5,000 a month. Every plan runs on the same managed infrastructure with the same security.

Sharing follows the repository, not the person. On Cloud Free nothing is shared. On Cloud Pro, the memory of a repository your team connects is read and written by every teammate’s agent, while a repository only you connect stays yours.

Yes. Start on Cloud Free and upgrade whenever you’re ready. Because it’s hosted, your memories are always there: nothing to export, import or migrate.

A few minutes. Run npx -y recallium install: it finds the coding agents on your machine, signs you in once and connects each one to Recallium Cloud. There is no server to run.

Your engineering memory belongs to you. We do not train AI models on your memories and we do not sell your data. Sharing follows the repository, every tenant is isolated by row-level security in the database, and the server decides who reads a project’s memory. How your data is handled is set out on the Security and trust page, in the privacy policy and in the Data Processing Addendum.