Persistent knowledge management for AI agents — store and retrieve project patterns, decisions, and architectural rules across sessions using the brv CLI.
by
byteroverinc
ByteRover provides long-term memory for AI agents through a local knowledge management system.
Using the brv CLI, your agent stores project patterns, architectural decisions, and coding rules as Markdown files in `.brv/context-tree/`, then queries them before tasks to keep work consistent across sessions.
Run `brv query` to retrieve relevant knowledge before starting a task.
After changes, run `brv curate` to save new patterns, decisions, and rules.
Knowledge is stored as human-readable Markdown in `.brv/context-tree/` for version control.
Use `brv login`, `brv push`, and `brv pull` for cloud sync when needed.
Retrieve relevant context before any task using `brv query` with natural language questions.
Save new patterns, decisions, and rules with `brv curate`, optionally including up to 5 source files.
The default ByteRover LLM provider works out of the box for query and curate operations.
Optionally connect OpenAI, Anthropic, Google, or other providers with your own API key.
Review past curations with `brv curate view`, and filter by time or status.
Push and pull knowledge to ByteRover cloud for team sharing (requires authentication).
All knowledge is stored as Markdown files in `.brv/context-tree/`.
Remember project conventions, coding patterns, and decisions across multiple sessions.
Store and recall architecture rules, auth patterns, database schemas, and design decisions.
Sync curated knowledge across teammates with optional cloud sync.
New teammates or AI agents can query the knowledge base to understand project conventions.
npx clawhub@latest install byteroverRun in your terminal, or click the Install button at the top of this page for one-click setup.
No. Local usage works without an API key; an API key is only needed for cloud sync features.
In the local `.brv/context-tree/` directory as human-readable Markdown files.
Yes. ByteRover supports bring-your-own LLM providers; follow the provider setup in the ByteRover docs.
Only if you enable cloud sync. Otherwise, data stays local in the context tree.
Use `brv query` to retrieve context and `brv curate` to save new knowledge.
Yes. Use `brv login` to authenticate, then `brv push` and `brv pull` to sync.