Feedback/Mar 26, 2026

Exportable chat history with markdown formatting

I need to copy‑paste conversation threads into documentation. An export button that creates a clean markdown file (including code blocks and timestamps) would streamline my workflow.

Pending
Feedback/Mar 26, 2026

Expanded language support for low‑resource languages

Please add high‑quality generation for Bengali, Swahili, and Ukrainian, with at least parity to English for summarization and translation tasks.

Pending
Feedback/Mar 26, 2026

Multi‑modal capability – include image input

I often need to annotate diagrams or extract data from screenshots. Adding native image upload and captioning would make Claude a one‑stop tool for visual analysis.

Pending
Feedback/Mar 26, 2026

Built‑in confidence scores or uncertainty indicators

When generating answers, a textual cue like “confidence: 85%” or a highlighted “low confidence” badge would let me instantaneously spot where I need to review or double‑check the output.

Pending
Feedback/Mar 26, 2026

Scheduled or batch inference API endpoints

I run nightly jobs that process thousands of customer support tickets. A scheduled or batch endpoint that accepts a list of prompts and returns responses in bulk would save us hours of orchestration.

Pending
Feedback/Mar 26, 2026

Custom temperature and top_p sliders in the UI

The web interface should expose fine‑grained controls for temperature and top_p (e.g., 0.0‑1.0 sliders) so I can reproduce deterministic outputs for testing and increase creativity on demand.

Pending
Feedback/Mar 26, 2026

Direct file upload support for multiple formats

I currently can only paste text. It would be extremely helpful to be able to upload Word docs, Excel spreadsheets, and PowerPoint presentations directly, with preserving tables and formatting for accurate extraction.

Pending
Feedback/Mar 26, 2026

Longer context window for deeper document analysis

I need to process PDFs up to 100 pages and maintain context across the entire file. Please increase the token limit to at least 250K tokens so I can upload full reports without manual chunking.

Pending