Starstruk Logo MarkStarstruk

v0.65.0

Mesh-LLM/mesh-llmv0.65.0May 1, 2026by github-actions[bot]

AI Summary

This major feature release adds a UI playground, per-model parallel configuration controls, and OpenAI endpoint passthrough plugins, alongside improvements to logging, metrics, and GPU handling.

Key Highlights

  • Introduced UI playground and completely revised console output.
  • Added per-model parallel configuration to control request slots.
  • Added OpenAI endpoint passthrough plugin support for vLLM, TGI, and Ollama.
  • Improved Vulkan device selection across startup and GPU reporting.

New Features

  • UI playground
  • per-model parallel configuration
  • OpenAI passthrough plugin
  • JSONL formatter for piped log output
  • metrics collector support

Full Release Notes

## What's Changed
* chore(ci): build SDK stuff when Rust / FFI changes by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/353
* Add per-model `parallel` config to control the number of parallel request slots per model by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/357
* Fix startup on nodes with pinned Tegra-based GPU SoCs by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/360
* Bump iroh to 0.98, move legacy Fly relay to tools/ by @michaelneale in https://github.com/Mesh-LLM/mesh-llm/pull/362
* Corrects duplicated log extension on logfiles by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/358
* refactor(cli): invert enumerate-host to opt-out with --no-enumerate-host by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/350
* chore(docs): update docs with directional metrics plan by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/351
* Fix passing parallel via config.toml by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/368
* Add support to output to multiple different log files by port for llamacpp servers by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/367
* Add mesh join age display with backend timestamp support by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/345
* api: expose read-only model search on the management API by @IvGolovach in https://github.com/Mesh-LLM/mesh-llm/pull/365
* Add delete option to model CLI flags to remove just a single managed model by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/366
* Add support to explicitly publish mesh via command-line by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/371
* api: add local explicit-interest submission for canonical model refs by @IvGolovach in https://github.com/Mesh-LLM/mesh-llm/pull/374
* Add persistent OpenCode config writing with `--write` flag by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/373
* paginate long model listings in interactive terminals by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/376
* Show full split GGUF sizes in model search results by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/377
* Fix: Add full workspace copy to Dockerfiles to resolve missing crates by @intenso in https://github.com/Mesh-LLM/mesh-llm/pull/363
* feature(ui-playground): add a UI playground and components by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/375
* chore(clippy): address clippy issues and warnings by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/378
* Fix Windows CUDA builds and dashboard VRAM display by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/381
* fix(models): handle HF snapshot/ref mismatch in updates flow by @josesantoscorte in https://github.com/Mesh-LLM/mesh-llm/pull/380
* Fix Vulkan device selection across startup and GPU reporting by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/383
* Build llama.cpp from an upstream patch queue by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/387
* Fix llama.cpp latest prepare workflow by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/391
* Add llama upstream pin canary by @i386 in https://github.com/Mesh-LLM/mesh-llm/pull/390
* openai-endpoint plugin: vLLM, TGI, Ollama passthrough by @michaelneale in https://github.com/Mesh-LLM/mesh-llm/pull/395
* feature(console-output): completely revise mesh-llm console output for JSONL output + pretty printing by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/388
* feature(metrics-collector): add support to centralize storage for met… by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/386
* Recovered peers can rejoin the mesh after dead_peers TTL expires by @michaelneale in https://github.com/Mesh-LLM/mesh-llm/pull/405
* Fix SIGABRT crash on Mac Studio (iroh 0.98 double-free in Endpoint::online) by @michaelneale in https://github.com/Mesh-LLM/mesh-llm/pull/407
* chore: add JSONL formatter for piped log output by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/408
* fix(tui): fix TUI interrupting error output for 'mesh-llm model show' by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/409
* feature(runtime): add model context info to runtime API by @ndizazzo in https://github.com/Mesh-LLM/mesh-llm/pull/410

## New Contributors
* @josesantoscorte made their first contribution in https://github.com/Mesh-LLM/mesh-llm/pull/380

**Full Changelog**: https://github.com/Mesh-LLM/mesh-llm/compare/v0.64.0...v0.65.0