ruby_llm_mesh v1.0.0 — stable documentation & brand release

ruby_llm_mesh v1.0.0 — stable documentation & brand release

Sovereign multi-provider AI mesh for Ruby & Rails

Highlights

• Official logo shipped at docs/logo.svg and featured in the README
• Badge pack with docs, release, license, status, version, category, and pages badges
• Detailed notes for integrators and diligence readers (CHANGELOG, SUPPORT, SECURITY, CONTRIBUTING)
• GitHub Pages documentation surface at https://theworker02.github.io/ruby_llm_mesh/
• Stable tag v1.0.0 as the portfolio baseline for this product

What changed in this release

1. Ensured brand mark and README presentation meet portfolio standards.
2. Expanded badge strip with explanatory notes table.
3. Added or refreshed diligence documentation.
4. Published narrative release notes for the 1.0 documentation line.

Verification

• README shows the official logo.
• Badges resolve via shields.io.
• CHANGELOG.md records the 1.0.0 documentation milestone.
• This GitHub Release tag is v1.0.0.

Upgrade / checkout

git clone https://github.com/theworker02/ruby_llm_mesh.git
cd ruby_llm_mesh
git checkout v1.0.0

Subsequent feature tags may advance beyond 1.0.0; this release remains the documented stable baseline for brand and docs completeness.

ruby_llm_mesh 2.2.0 — Provider retries

max_retries now retries transient TimeoutError, RateLimitError, and 5xx/transport ProviderErrors with exponential backoff before falling over.

• retry_backoff (default 0.1s, env RUBY_LLM_MESH_RETRY_BACKOFF) doubles after each failed attempt
• Authentication failures are never retried

Published on RubyGems: https://rubygems.org/gems/ruby_llm_mesh/versions/2.2.0

What's Changed

• feat: ruby_llm_mesh 2.2.0 — retry transient provider failures by @theworker02 in #2

Full Changelog: v2.1.0...v2.2.0

v2.0.0 — Sovereign Mesh Core & Native FFI

Summary

Major release: sovereign mesh with optional Rust chimera_core FFI, strategy-based execute(), real cloud failover, plus optional semantic cache and peer health.

Highlights

• Native FFI core (ext/chimera_core/) — start_node, node_alive, execute_wasm_payload with pure-Ruby fallback when uncompiled
• AiAgentRouter.execute(intent:, strategy:) — :auto, :p2p_mesh, :cloud, and per-provider strategies
• Cloud paths remain real HTTP (OpenAI / Anthropic / local_node), not stubs
• Optional semantic cache (memory / Redis) and local peer health monitoring
• Trusted publishing workflow consolidated to .github/workflows/release.yml

Breaking changes (0.1.0 → 2.0.0)

• Major version bump — pins of ~> 0.1 will not upgrade automatically
• New defaults: auto_boot_mesh: true, mesh_port: 4233 (affect :auto / :p2p_mesh only)
• complete API shape preserved; semantic cache remains off by default
• Trusted publisher workflow filename is now release.yml (was push_gem.yml)

Install

gem install ruby_llm_mesh
# optional native core:
bundle exec rake compile   # requires Rust/cargo

Official gem: https://rubygems.org/gems/ruby_llm_mesh

RubyGems trusted publishing

If the release workflow fails on RubyGems, configure a pending trusted publisher:

Field	Value
Gem name	ruby_llm_mesh
Owner	theworker02
Repository	ruby_llm_mesh
Workflow	release.yml
Environment	release

Also ensure a GitHub Environment named release exists on the repo.

v0.1.0 — Initial release

ruby_llm_mesh 0.1.0

Initial release of unified multi-provider AI routing for Ruby and Rails.

Highlights

• Multi-provider routing — Call OpenAI, Anthropic, and OpenAI-compatible local nodes (Ollama, LM Studio, vLLM, etc.) through one DSL (AiAgentRouter.complete / RubyLlmMesh.complete)
• Circuit breaker & fallback — Per-provider circuits with configurable failure threshold and reset timeout; automatic fall-through across the provider ladder when a node fails or its circuit is open
• RAG helpers — Text chunker, bag-of-words embeddings for prototyping, and tool schema adapters for OpenAI-style tool definitions
• acts_as_ai_agent — Optional ActiveRecord mixin for conversational memory, a simple semantic cache, and audit logging (Rails railtie loads only when Rails is present)
• Docs site — Project documentation on GitHub Pages
• Trusted publishing — GitHub Actions workflow (.github/workflows/push_gem.yml) for RubyGems OIDC trusted publishing on v* tags

Install

gem install ruby_llm_mesh

RubyGems: https://rubygems.org/gems/ruby_llm_mesh

Docs: https://theworker02.github.io/ruby_llm_mesh/

See CHANGELOG.md for the full list of changes.