CLI Reference¶

All commands are under basemode. Running basemode with no explicit subcommand defaults to run.

Generation¶

Generate continuation text (single stream or parallel branches).

basemode [PREFIX] [OPTIONS]

Option	Default	Description
`-m`, `--model`	stored default or `gpt-4o-mini`	Model to use
`-n`, `--branches`	`1`	Number of parallel continuations
`-M`, `--max-tokens`	`200`	Max output tokens
`-t`, `--temperature`	`0.9`	Sampling temperature (when model allows)
`-s`, `--strategy`	auto	Force strategy selection
`--rewind`	`false`	Rewind short trailing fragments before generation
`--show-strategy`	`false`	Print selected strategy
`--show-usage`	`false`	Print token estimate after generation
`--show-cost`	`false`	Print estimated cost after generation

PREFIX can come from stdin when omitted.

List LiteLLM-known models.

basemode models [--provider openai] [--search claude] [--available] [--verified] [--json]

--available limits to providers with configured keys.
--verified limits to models tracked in the verified-models registry.
--json emits structured picker metadata (provider, availability, reliability, pricing fields when known).

List all known providers.

basemode providers

List supported continuation strategies.

basemode strategies

Show normalized model ID, selected strategy, token limits, and pricing metadata.

basemode info claude-sonnet-4-6

Manage persisted API keys (~/.config/basemode/auth.json).

basemode keys set openai
basemode keys list
basemode keys get anthropic

Show/set/unset default model.

basemode default
basemode default gpt-4o-mini
basemode default --unset