## for Gemini
Long-term memory for Gemini CLI
A persistent knowledge graph the CLI reads and writes across every session.
The CLI starts every session cold
Architecture decisions, conventions, and yesterday's context don't carry forward unless you paste them in again.
- —Gemini CLI has no built-in long-term memory. Every session is a fresh start.
- —You re-explain architecture, conventions, and project history on each run.
- —Context built up in Gemini can't follow you to Claude, ChatGPT, or your editor.
- —Notes scattered across chat history are impossible to search later.
Basic Memory changes that
Give Gemini a persistent, searchable knowledge base it can read and write to.
Persistent context across sessions
Gemini reads and writes a knowledge base that survives the session. Architecture decisions, conventions, and accumulated project knowledge always on hand.
Full-text, semantic, and graph search
Every session builds a searchable knowledge graph. Gemini recalls by keyword, by meaning, or by walking the relations between notes.
Plain Markdown you own
Notes are files on your machine. Read them in any editor, version them in Git, take them anywhere. Open the same vault in Obsidian alongside Gemini if you want.
Native MCP server
Connects over the Model Context Protocol — the same standard Gemini CLI uses for tools. The same knowledge graph also works in Claude, Cursor, Codex, and any MCP client.
One command to get started
Quick Setup
bm cloud login && bm mcp install --client geminiOr add it by hand. In ~/.gemini/settings.json: `"mcpServers": { "basic-memory": { "command": "bm", "args": ["mcp"] } }`. The bm CLI is the Basic Memory command-line tool (`pip install basic-memory` or `brew install basic-memory`). Run `gemini` and the new tools are available. Connect to Basic Memory Cloud or a local project — your choice. Full setup guide →
Start building with Gemini
Try Basic Memory free for 7 days. Your notes stay yours either way.
7-Day Free Trial
$15/seat/month
Graduated team pricing- AI collaboration via MCP
- Shared workspaces for teams and agents
- Team member management and audit logs
- Full-text search across notes and projects
- Desktop, mobile, and web access
- Private, exportable Markdown files
- 50,000 notes per seat
- 1,000 note updates per seat per day
- Cancel anytime, your data stays yours
Open source available for power users
View on GitHubFrequently asked questions
- How do I connect Basic Memory to the Gemini CLI?
- Two ways. Easy: run
bm cloud login && bm mcp install --client geminiand you’re done. Manual: add an entry to~/.gemini/settings.json:"mcpServers": { "basic-memory": { "command": "bm", "args": ["mcp"] } }. ThebmCLI is the Basic Memory command-line tool (install withpip install basic-memoryorbrew install basic-memory). Rungeminiand the new tools are available. - Does the Gemini CLI have built-in long-term memory?
- No. The CLI starts every session cold — architecture, conventions, and decisions from yesterday don’t carry forward unless you paste them in. Basic Memory is the long-term layer that fixes that, queryable by the agent via MCP.
- What does Gemini actually get?
- Through MCP, Gemini gets full-text and semantic search across your notes, read and write any note by title or permalink, navigate the knowledge graph via memory:// URLs, list directories, get recent activity, switch between projects, and validate notes against schemas. The same tool surface every other MCP client gets.
- Does this work alongside other AI tools?
- Yes. Connect the same knowledge base to Claude, ChatGPT, Cursor, Codex, Hermes, OpenClaw, Obsidian, VS Code — anything that speaks MCP. Notes you write in Gemini are immediately available everywhere else.
- What does it cost with Gemini?
- Local Basic Memory is free and open source (AGPL-3.0) with no API costs. Basic Memory Cloud, for synced and hosted access plus team workspaces, starts at $15/seat/month with a 7-day trial. Gemini itself is billed by Google separately.
