Citiview Context Packer

Choose this endpoint when you need to fit relevant content from long documents or multiple web pages into a limited token window for an LLM prompt. It is especially useful when you have more source material than your context window can hold and need automatic relevance ranking and source attribution. Prefer it over full document retrieval when token efficiency matters, or over generic summarization when you need question-specific passage selection rather than a general summary.

• Missing 'question' parameter returns a validation error
• Unreachable or invalid URLs result in fetch errors for those sources
• Token budget set too low may return very few or truncated passages
• Overly broad questions may yield low-relevance passages
• Large numbers of long texts may increase latency or hit size limits
• Malformed URL format in the pipe-separated list causes parsing errors