Custom Model (Claude Message API) disables Edit + Copy returns empty content

Hi TypingMind team — I’m encountering a serious UX bug specifically with Custom Models configured using Claude Message API (Anthropic /v1/messages). TypingMind supports adding custom models and working across OpenAI/Claude/Gemini providers.

Summary

When I add a Custom Model and select different API Type options, the message actions behave inconsistently:

Case 1 — API Type: OpenAI Chat Completions API (/v1/chat/completions)

Everything works normally:

• Edit (both user message + assistant message) works

• Regenerate works

• Copy works (copies the actual content)

(OpenAI Chat Completions endpoints are part of TypingMind’s supported endpoint setup.)

Case 2 — API Type: Claude Message API (/v1/messages)

Two issues happen consistently:

1. Edit button is disabled (greyed out) and cannot be clicked

2. After the AI generates a response, Copy produces empty content (clipboard ends up blank). I have to manually select all text to copy.

TypingMind’s docs indicate that editing/regenerating should be supported as part of the chat workflow (Chat Thread preserves versions when you regenerate or edit).
TypingMind also documents copy functionality (e.g., “Copy Last Message” shortcut).
And “Claude Message API” is explicitly an API Type used for custom model setup.

Steps to reproduce

1. Go to Models → Add Custom Model

2. Create a model with API Type = Claude Message API

3. Start a new chat with this model and send any message

4. After the response appears:

   • Try clicking Edit on a message → button is greyed out

   • Click Copy on the assistant response → clipboard becomes empty (no content)

Expected behavior

Editing/regenerating should work normally (and create a new thread/version), and Copy should copy the full message content, consistent with TypingMind’s chat thread and copy features.

Actual behavior

• Edit disabled (unclickable)

• Copy returns empty content only for Claude Message API custom models

Impact / severity

This is very disruptive: I can’t iteratively refine prompts by editing messages, and copying outputs becomes tedious and error-prone.

Additional note

Under the same “edit/copy/regenerate” workflow, I do not see these issues in other third-party LLM clients (e.g., Cherry Studio), so it seems TypingMind-specific.

Thanks — this one feels urgent because it breaks core chat operations for Claude-based custom models.