Skip to content

basemode

basemode is a Python library and CLI for clean text continuation with any LLM.

The problem it solves: modern LLMs are chat-tuned and respond to continuation prompts with preambles like "Sure! Here's a continuation:" rather than just writing the next words. basemode wraps any LLM with provider-specific coercion strategies that force it to output raw continuation text — no acknowledgment, no formatting, no wrapper.

It is the model layer underneath index (basemode-loom). loom handles trees, storage, and navigation; basemode handles the actual generation.

What it provides

  • continue_text() — stream a single continuation from any model
  • branch_text() — stream N parallel continuations simultaneously
  • Automatic strategy detection — picks the right coercion technique per model
  • Token boundary healing — fixes split words and spacing artifacts during streaming
  • 20+ provider support — OpenAI, Anthropic, Google, Groq, Together, OpenRouter, and more
  • CLI — basemode "your text" for one-shot use

Quick example

import asyncio
from basemode import continue_text

async def main():
    async for token in continue_text(
        "The ship rounded the headland and",
        model="gpt-4o-mini",
        max_tokens=200,
    ):
        print(token, end="", flush=True)

asyncio.run(main())

# CLI
basemode "The ship rounded the headland and" --model gpt-4o-mini