byterover-cli

ByteRover CLI (the command brv) is a tool that gives AI coding assistants a persistent memory layer. When you use AI tools like Claude Code, Cursor, or similar coding assistants, they normally start fresh with each session and have no memory of what they learned about your project before. ByteRover addresses this by letting you store and retrieve structured knowledge about your codebase, which the AI can then use across sessions.

You run brv inside any project directory to start an interactive chat session powered by a language model of your choice. The tool can read and write files, run code, and save notes about your project to a searchable knowledge store called a context tree. In later sessions, the AI can query this stored knowledge to pick up where it left off, rather than rediscovering context from scratch.

The knowledge store supports version control that mirrors the concepts in git: you can branch, commit, push, pull, and merge your stored context just as you would source code. A cloud sync option lets teams share the same context tree across machines and teammates, so everyone's AI assistant works from the same base of project knowledge. A web dashboard is included for browsing and editing stored context without using the command line.

The CLI supports more than 20 language model providers including Anthropic, OpenAI, Google, Groq, and others, so you are not tied to a single AI service. It also works as an MCP (Model Context Protocol) server, which is a standard way for AI tools to communicate with external data sources.

The project includes benchmark results on two long-term memory evaluation datasets, reporting above 92% accuracy. It installs via npm or a standalone shell script on macOS and Linux. The Elastic 2.0 license allows personal and most commercial use; check the full terms if you plan to redistribute the software or offer it as a hosted service.