Long documents, read whole.

Give the model the whole thing: a contract bundle, a stack of papers, a codebase, a year of meeting notes. Ask questions across all of it instead of pasting in pieces.

Pre-launch. Accounts are open; the API and checkout open soon.

How it runs.

  1. 01

    Put it all in.

    Up to the model’s full window: 1M tokens on Kimi K3.

  2. 02

    Ask across it.

    Compare clauses, trace a function through the codebase, find what changed between drafts.

  3. 03

    Ask again.

    Every follow-up resends the whole document. On lanes that doesn’t cost more.

Why lanes fit it.

  • Input is the whole cost.

    Reading a million tokens to answer one question is what per-token bills charge most for.

  • Long context has a lane price.

    A request past the default context takes twice its lanes on Solo and Pro, and nothing extra on the others. No surcharge per token.

  • Keep the conversation going.

    Ask as many follow-ups as you like.

OpenAI Python

python
import osfrom openai import OpenAI client = OpenAI(    base_url="https://api.voidstone.net/v1",    api_key=os.environ["VOIDSTONE_API_KEY"],) stream = client.chat.completions.create(    model="kimi-k3",    messages=[{"role": "user", "content": "Explain this stack trace."}],    stream=True,)for chunk in stream:    print(chunk.choices[0].delta.content or "", end="")