onesentence.fun

Any question.
One sentence.

Paste as much context as you like, ask anything, and GPT-6 Astra Pro reasons through all of it at maximum effort, then answers in exactly one sentence.

0 / 390,000 charactersEnter sends · Shift+Enter for a new line

The system prompt

sent verbatim with every question
You are onesentence.fun. Read and reason carefully through ALL of the information presented, every word of the context and the question, no matter how long it is, before you answer. Weigh the details, resolve contradictions, and use everything that is relevant. Then give your answer in exactly one sentence: a single, complete, grammatical sentence that ends with one terminal punctuation mark. No preamble, no lists, no line breaks, no follow-up questions, and never a second sentence. If the request is ambiguous, answer its most likely intent; if it cannot be answered from the information given and what you know, say so in one sentence.
model
openai/gpt-6-astra-pro
reasoning effort
max
max output
128,000 tokens
max input
390,000 chars

How it works

  1. Sign in with your phone number (a one-time text-message code, via Privy).
  2. Paste your context and question, up to 390,000 characters.
  3. The question goes to openai/gpt-6-astra-pro through OpenRouter with reasoning effort max and the full 128,000-token output budget.
  4. You get one sentence back. If the model ever writes more, the server keeps only the first sentence.

Pricing: answers are free to you; each one shows the real cost billed by OpenRouter ($10 per 1M input tokens, $50 per 1M output and reasoning tokens). Limits: one question at a time, 10 per hour, 30 per day.

Agents and developers: see the API docs, OpenAPI spec, llms.txt, and the MCP server at https://onesentence.fun/mcp.