CLI Reference¶
The basemode CLI is invoked as basemode. The default command (basemode "text") is shorthand for basemode run "text".
basemode / basemode run¶
Generate a continuation.
| Option | Default | Description |
|---|---|---|
--model, -m |
stored default or gpt-4o-mini |
Model to use |
--max-tokens, -M |
200 |
Max tokens to generate |
--temperature, -t |
0.9 |
Sampling temperature |
--branches, -n |
1 |
Number of parallel continuations |
--strategy, -s |
auto | Force a specific strategy (system, prefill, few_shot, completion, fim) |
--context, -c |
"" |
System context / framing |
--rewind |
off | Rewind prefix to word boundary before sending |
--show-strategy |
off | Print the detected strategy and exit |
# Pipe text in
cat chapter.txt | basemode --max-tokens 500
# Parallel branches, shown side-by-side
basemode "She opened the door to find" --branches 4
# Show what strategy will be used without generating
basemode "prefix" --show-strategy
basemode models¶
List available models.
basemode models
basemode models --available # only models with API keys configured
basemode models --provider openai # filter by provider
basemode models --search claude # search by name fragment
basemode providers¶
List all known providers.
basemode strategies¶
List all available continuation strategies with descriptions.
basemode info¶
Show strategy detection and pricing info for a model.
Outputs: - Detected strategy - Input/output cost per 1M tokens - Context and output token limits
basemode keys¶
Manage stored API keys.
basemode keys set openai # prompts for key value
basemode keys set anthropic
basemode keys get openai # show stored key (masked)
basemode keys list # show all stored keys (masked)
Keys are stored in ~/.config/basemode/auth.json with 0o600 permissions.
basemode default¶
Get or set the default model.
basemode default # show current default
basemode default gpt-4o-mini # set default
basemode default --clear # clear default
Once set, --model can be omitted from basemode run.