🤖 Re-releasing the Claude LLM Pack with support for Claude 4, cost tracking, and more!

Hey Coda Makers!

Just released a major update to the Claude AI Pack

Here’s what’s new:

  • Claude 4 Support: Claude Opus 4 (32K output!), Claude Sonnet 4 (64K output!), plus smart aliases that auto-update to the latest versions.
  • Cost Tracking: Not just token tracking - includes detailed breakdown of input cost, output cost, total cost. Works even when JSON parsing fails.
  • Enhanced JSON Handling: Aggressive sanitization removes control characters that break parsing. Dual-attempt parsing (original + sanitized). Detailed error reporting with both attempts included. Token preservation even when Claude returns invalid JSON
  • Extended Thinking Improvements: Multi-model support across Opus 4, Sonnet 4, and Sonnet 3.7. Future-proofed for whatever Claude ships next.
  • Zero breaking changes: Your existing formulas work as-is, just way more reliable now!

Anyone building cool automation workflows with this? Would love to see what you’re creating :+1:

Cheers,
Jon

7 Likes