ai_assist¶
CLI tool to manage AI Assistant workspaces.
Usage:
Options:
--install-completion: Install completion for the current shell.--show-completion: Show completion for the current shell, to copy it or customize the installation.--help: Show this message and exit.
Commands:
init: Initialize a new AI Assistant workspace.run: Start backend and frontend services.workspace: Workspace management commandsglobal: Global resources managementknowledge: Knowledge base managementnote: Customer note commands
ai_assist init¶
Initialize a new AI Assistant workspace.
Creates: - ~/.ai_assist/: Global directory for cross-workspace resources - prompts/: Shared prompt templates - agents/: Agent definitions - tools/: Shared tool definitions - models/: Model configurations - cache/: Shared cache
- <workspace>/: Local workspace directory
- data/chroma/: Vector database
- prompts/: Local prompt templates
- tools/: Local tool definitions
- history/: Chat history
- summaries/: Conversation summaries
- notes/: Documents for RAG
Usage:
Arguments:
[PATH]: Path where to create the workspace. Defaults to current directory.
Options:
-n, --name TEXT: Workspace name. Defaults to directory name.-f, --force: Overwrite existing workspace configuration.--help: Show this message and exit.
ai_assist run¶
Start backend and frontend services.
By default, uses Docker Compose to start services with pre-built images. Use --dev flag to run services directly using uv and npm (development mode).
The command will auto-detect the workspace by looking for the workspace marker in the current directory or parent directories. You can also specify a workspace path.
Usage:
Arguments:
[WORKSPACE_PATH]: Path to workspace. Defaults to current directory (auto-detected).
Options:
-d, --dev: Run in development mode (uses uv and npm instead of Docker).--help: Show this message and exit.
ai_assist workspace¶
Workspace management commands
Usage:
Options:
--help: Show this message and exit.
Commands:
status: Show workspace status and directory summary.clean: Clean workspace data (history, summaries,...list: List all known workspaces.
ai_assist workspace status¶
Show workspace status and directory summary.
Usage:
Arguments:
[PATH]: Path to workspace. Defaults to current directory.
Options:
--help: Show this message and exit.
ai_assist workspace clean¶
Clean workspace data (history, summaries, cache).
Usage:
Arguments:
[PATH]: Path to workspace. Defaults to current directory.
Options:
-y, --yes: Skip confirmation prompt.--help: Show this message and exit.
ai_assist workspace list¶
List all known workspaces.
Usage:
Options:
--help: Show this message and exit.
ai_assist global¶
Global resources management
Usage:
Options:
--help: Show this message and exit.
Commands:
status: Show global AI Assist home status.config: Display global configuration file.prompts: List available global prompts.agents: List available global agents.tools: List available global tools.tree: Show the global directory structure.
ai_assist global status¶
Show global AI Assist home status.
Usage:
Options:
--help: Show this message and exit.
ai_assist global config¶
Display global configuration file.
Usage:
Options:
--help: Show this message and exit.
ai_assist global prompts¶
List available global prompts.
Usage:
Options:
--help: Show this message and exit.
ai_assist global agents¶
List available global agents.
Usage:
Options:
--help: Show this message and exit.
ai_assist global tools¶
List available global tools.
Usage:
Options:
--help: Show this message and exit.
ai_assist global tree¶
Show the global directory structure.
Usage:
Options:
--help: Show this message and exit.
ai_assist knowledge¶
Knowledge base management
Usage:
Options:
--help: Show this message and exit.
Commands:
process: Process knowledge documents from a JSON...stats: Show RAG vector store statistics.
ai_assist knowledge process¶
Process knowledge documents from a JSON specification file.
The JSON file should contain an array of document specifications:
[ {"document_type": "website", "uri": "https://...", "collection": "flink"}, {"document_type": "folder", "uri": "$HOME/docs", "collection": "python"} ]
Supported document_type values: website, folder, markdown
The 'collection' field is mapped to 'category' in the backend for filtering. Environment variables ($HOME) and ~ are expanded in URIs.
Usage:
Arguments:
JSON_FILE: Path to JSON file containing document specifications. [required]
Options:
-b, --backend-url TEXT: Backend API URL. Defaults to AI_ASSIST_BACKEND_URL env var or http://localhost:8000/api-n, --dry-run: Validate JSON and show what would be processed without making API calls.-f, --force: Force re-indexing even if content unchanged.-v, --verbose: Show detailed output including errors.--help: Show this message and exit.
ai_assist knowledge stats¶
Show RAG vector store statistics.
Usage:
Options:
-b, --backend-url TEXT: Backend API URL. Defaults to AI_ASSIST_BACKEND_URL env var or http://localhost:8000/api--help: Show this message and exit.
ai_assist note¶
Customer note commands
Usage:
Options:
--help: Show this message and exit.
Commands:
parse: Parse a customer note markdown file and...
ai_assist note parse¶
Parse a customer note markdown file and create organization, project, persons, and meeting refs.
The file should follow a structure like: H1 = org name, ## Team / ## <Org> = people,
Products/Context/Technology stack = context, ## Past steps/Next steps = steps,¶
Discovery call/Questions = meeting sections. Uses an LLM to extract structured¶
data and optionally creates entities via the backend API.
Usage:
Arguments:
FILE_PATH: Path to customer note markdown file (e.g. index.md). [required]
Options:
-b, --backend-url TEXT: Backend API URL. Defaults to AI_ASSIST_BACKEND_URL or http://localhost:8000/api-n, --dry-run: Only parse and show extracted data; do not call the backend.-v, --verbose: Show detailed output.--help: Show this message and exit.