v0.6.1

Vexa-ai/vexav0.6.1Sep 25, 2026by Pringled

Upgrading?

AI Summary

Added idle TTL for MCP in-memory index cache and support for new AI model installers (Grok Build, Qwen Code, Cline, and Kilo Code).

Key Highlights

  • Added idle TTL for MCP in-memory index cache.
  • Added Grok Build, Qwen Code, Cline, and Kilo Code installer support.
  • Simplified SEMBLE_MAX_FILE_BYTES handling and CLI logging setup.
  • Added zvec-grep to benchmarks.

New Features

  • MCP cache management improvements.
  • New AI model installer support.
  • Benchmark updates.

Full Release Notes

## What's Changed
* benchmarks: Add zvec-grep to benchmarks by @Pringled in https://github.com/MinishLab/semble/pull/268
* chore: Hide HF unauthenticated warning by @Pringled in https://github.com/MinishLab/semble/pull/270
* chore: Simplify SEMBLE_MAX_FILE_BYTES handling and CLI logging setup by @Pringled in https://github.com/MinishLab/semble/pull/271
* chore: Remove dead code and deduplicate internals by @Pringled in https://github.com/MinishLab/semble/pull/272
* feat: Add idle TTL for MCP in-memory index cache by @Pringled in https://github.com/MinishLab/semble/pull/273
* feat: Add Grok Build, Qwen Code, Cline, and Kilo Code installer support by @Pringled in https://github.com/MinishLab/semble/pull/275


**Full Changelog**: https://github.com/MinishLab/semble/compare/v0.6.0...v0.6.1

Related Repositories