CLI Reference¶
All commands are under basemode-loom. Run basemode-loom --help or basemode-loom <command> --help for up-to-date flags.
Generation¶
run¶
Generate continuations from a new prompt.
| Option | Default | Description |
|---|---|---|
-m, --model |
gpt-4o-mini |
Model to use |
-M, --max-tokens |
200 |
Max tokens per continuation |
-n, --n-branches |
3 |
Number of parallel continuations |
-t, --temperature |
0.9 |
Sampling temperature |
continue¶
Generate continuations from the current active node.
Uses the same model/settings as the active node's tree unless overridden.
| Option | Description |
|---|---|
-b, --branch |
Branch index to continue from (default: checked-out child) |
-n, --n-branches |
Number of continuations |
Navigation¶
view¶
Open the interactive TUI explorer.
If given a node ID or file path, opens that tree. Otherwise opens the most recently active tree. See TUI Guide for keybindings.
nodes¶
List recent nodes.
active¶
Show the current active node's full text and metadata.
show¶
Show a specific node.
Supports prefix matching — you can use the first few characters of a node ID.
children¶
List the children of a node.
Analysis¶
stats¶
Show quantitative statistics for a tree.
Includes tree depth, expansion rates, model breakdown, and descendant scores. Pass --json for machine-readable output.
Import / Export¶
export¶
Export a tree to a file or stdout.
basemode-loom export [NODE_ID] --to json
basemode-loom export [NODE_ID] --to file.md
basemode-loom export [NODE_ID] --to output.json
Supported formats: json, md (markdown), or a file path with an appropriate extension.
import¶
Import a tree from a file. Supports several legacy formats (tinyloom, minihf, bonsai, basemode-json).
Server¶
serve¶
Start the FastAPI web server.
See Web Server for API documentation.