Feature request: parameter max_output_tokens

[MatejRojec]

Would it be possible to add a parameter to the Jina Reader API that allows limiting the output length (e.g., max_output_tokens). This would let users retrieve partial content, such as the first 5,000 tokens, without hitting the token budget errors or receiving full responses. It would improve efficiency for workflows that only need the initial portion of extracted text.

[AyusKumarPathak]

@MatejRojec @joshdevins @nomagick @hanxiao @crazyurus
Hello maintainers 👋

I want to implement support for the max_output_tokens parameter.
Solution plan:

1. Add the new parameter in the API schema.
2. Pass it downstream to the LLM inference call.
3. Update tests to verify output truncation behavior.
4. Update README/examples with usage.

Please assign this to me if that’s okay.

— AyusKumarPathak