DecisionNode

# DecisionNode > Structured, queryable memory for development decisions. Stores architectural choices as vector embeddings, exposes them to AI agents via MCP. ## Quick Install ```bash npm install -g decisionnode cd your-project decide init # creates project store + .mcp.json for AI clients decide setup # configure Gemini API key (free tier) decide add # add your first decision ``` ## Three interfaces - **CLI (`decide` or `decisionnode`)** — for you. Add, search, list, edit, deprecate, export/import, configure. - **MCP server (`decide-mcp`)** — for your AI. AI clients connect via MCP and read/write decisions through 9 tools. - **Web UI (`decide ui`)** — local visualization. Three views (graph / vector space / list) with a live MCP pulse that animates AI activity in real time. Read-only; CLI and MCP remain the write paths. `decide ui -d` runs it in the background. ## Documentation - [Introduction](https://decisionnode.dev/docs) - [Quickstart](https://decisionnode.dev/docs/quickstart) - [Installation](https://decisionnode.dev/docs/installation) - [Configuration](https://decisionnode.dev/docs/setup) - [Decision Nodes](https://decisionnode.dev/docs/decisions) — structure, fields, lifecycle - [Context Engine](https://decisionnode.dev/docs/context) — embedding, search, conflict detection - [CLI Reference](https://decisionnode.dev/docs/cli) — all commands - [MCP Server](https://decisionnode.dev/docs/mcp) — all 9 tools, setup - [Workflows](https://decisionnode.dev/docs/workflows) — common patterns ## Full docs (single file) For complete documentation as a single markdown file: https://decisionnode.dev/decisionnode-docs.md ## Links - GitHub: https://github.com/decisionnode/decisionnode - License: MIT