llm.rb 4.7.0 → 4.9.0

data/lib/llm.rb

@@ -6,10 +6,13 @@ module LLM
   require_relative "llm/tracer"
   require_relative "llm/error"
   require_relative "llm/contract"
+  require_relative "llm/registry"
+  require_relative "llm/cost"
   require_relative "llm/usage"
   require_relative "llm/prompt"
   require_relative "llm/schema"
   require_relative "llm/object"
+  require_relative "llm/model"
   require_relative "llm/version"
   require_relative "llm/utils"
   require_relative "llm/message"
@@ -18,7 +21,7 @@ module LLM
   require_relative "llm/multipart"
   require_relative "llm/file"
   require_relative "llm/provider"
-  require_relative "llm/session"
+  require_relative "llm/context"
   require_relative "llm/agent"
   require_relative "llm/buffer"
   require_relative "llm/function"
@@ -29,7 +32,22 @@ module LLM
 
   ##
   # Thread-safe monitors for different contexts
-  @monitors = {require: Monitor.new, clients: Monitor.new, inherited: Monitor.new}
+  @monitors = {require: Monitor.new, clients: Monitor.new, inherited: Monitor.new, registry: Monitor.new}
+
+  ##
+  # Model registry
+  @registry = {}
+
+  ##
+  # @param [Symbol, LLM::Provider] llm
+  #  The name of a provider, or an instance of LLM::Provider
+  # @return [LLM::Object]
+  def self.registry_for(llm)
+    lock(:registry) do
+      name = Symbol === llm ? llm : llm.name
+      @registry[name] ||= Registry.for(name)
+    end
+  end
 
   module_function
 
@@ -75,10 +93,10 @@ module LLM
 
   ##
   # @param (see LLM::Provider#initialize)
-  # @return (see LLM::Gemini#initialize)
-  def gemini(**)
-    lock(:require) { require_relative "llm/providers/gemini" unless defined?(LLM::Gemini) }
-    LLM::Gemini.new(**)
+  # @return (see LLM::Google#initialize)
+  def google(**)
+    lock(:require) { require_relative "llm/providers/google" unless defined?(LLM::Google) }
+    LLM::Google.new(**)
   end
 
   ##
@@ -131,6 +149,22 @@ module LLM
     LLM::ZAI.new(**)
   end
 
+  ##
+  # @param [LLM::Provider, nil] llm
+  #  The provider to use for MCP transports that need one
+  # @param [Hash, nil] stdio
+  # @option stdio [Array<String>] :argv
+  #  The command to run for the MCP process
+  # @option stdio [Hash] :env
+  #  The environment variables to set for the MCP process
+  # @option stdio [String, nil] :cwd
+  #  The working directory for the MCP process
+  # @return [LLM::MCP]
+  def mcp(llm = nil, **)
+    lock(:require) { require_relative "llm/mcp" unless defined?(LLM::MCP) }
+    LLM::MCP.new(llm, **)
+  end
+
   ##
   # Define a function
   # @example

data/llm.gemspec

@@ -5,17 +5,52 @@ require_relative "lib/llm/version"
 Gem::Specification.new do |spec|
   spec.name = "llm.rb"
   spec.version = LLM::VERSION
-  spec.authors = ["Antar Azri", "0x1eef"]
-  spec.email = ["azantar@proton.me", "0x1eef@proton.me"]
+  spec.authors = ["Antar Azri", "0x1eef", "Christos Maris", "Rodrigo Serrano"]
+  spec.email = ["azantar@proton.me", "0x1eef@hardenedbsd.org"]
 
   spec.summary = <<~SUMMARY
-  llm.rb is a zero-dependency Ruby toolkit for Large Language Models that
-  includes OpenAI, Gemini, Anthropic, xAI (grok), zAI, DeepSeek, Ollama, and
-  LlamaCpp. The toolkit includes full support for chat, streaming, tool calling,
-  audio, images, files, and structured outputs.
+  llm.rb is a Ruby-centric toolkit for building real LLM-powered systems — where
+  LLMs are part of your architecture, not just API calls. It gives you explicit
+  control over contexts, tools, concurrency, and providers, so you can compose
+  reliable, production-ready workflows without hidden abstractions.
   SUMMARY
 
-  spec.description = spec.summary
+  spec.description = <<~DESCRIPTION
+  llm.rb is a Ruby-centric toolkit for building real LLM-powered systems — where
+  LLMs are part of your architecture, not just API calls. It gives you explicit
+  control over contexts, tools, concurrency, and providers, so you can compose
+  reliable, production-ready workflows without hidden abstractions.
+
+  Built for engineers who want to understand and control their LLM systems. No
+  frameworks, no hidden magic — just composable primitives for building real
+  applications, from scripts to full systems like Relay.
+
+  ## Key Features
+
+  - **Contexts are central** — Hold history, tools, schema, usage, cost, persistence, and execution state
+  - **Tool execution is explicit** — Run local, provider-native, and MCP tools sequentially or concurrently
+  - **One API across providers** — Unified interface for OpenAI, Anthropic, Google, xAI, zAI, DeepSeek, Ollama, and LlamaCpp
+  - **Thread-safe where it matters** — Providers are shareable, while contexts stay isolated and stateful
+  - **Production-ready** — Cost tracking, observability, persistence, and performance tuning built in
+  - **Stdlib-only by default** — Runs on Ruby standard library, with optional features loaded only when used
+
+  ## Capabilities
+
+  - Chat & Contexts with persistence
+  - Streaming responses
+  - Tool calling with JSON Schema validation
+  - Concurrent execution (threads, fibers, async tasks)
+  - Agents with auto-execution
+  - Structured outputs
+  - MCP (Model Context Protocol) support
+  - Multimodal inputs (text, images, audio, documents)
+  - Audio generation, transcription, translation
+  - Image generation and editing
+  - Files API for document processing
+  - Embeddings and vector stores
+  - Local model registry for capabilities, limits, and pricing
+  DESCRIPTION
+
   spec.license = "0BSD"
   spec.required_ruby_version = ">= 3.2.0"
 
@@ -23,10 +58,12 @@ Gem::Specification.new do |spec|
   spec.metadata["homepage_uri"] = "https://github.com/llmrb/llm.rb"
   spec.metadata["source_code_uri"] = "https://github.com/llmrb/llm.rb"
   spec.metadata["documentation_uri"] = "https://0x1eef.github.io/x/llm.rb"
+  spec.metadata["changelog_uri"] = "https://0x1eef.github.io/x/llm.rb/file.CHANGELOG.html"
 
   spec.files = Dir[
     "README.md", "LICENSE",
     "lib/*.rb", "lib/**/*.rb",
+    "data/*.json",
     "llm.gemspec"
   ]
   spec.require_paths = ["lib"]
@@ -44,4 +81,4 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "net-http-persistent", "~> 4.0"
   spec.add_development_dependency "opentelemetry-sdk", "~> 1.10"
   spec.add_development_dependency "logger", "~> 1.7"
-end
+end

metadata

@@ -1,11 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: llm.rb
 version: !ruby/object:Gem::Version
-  version: 4.7.0
+  version: 4.9.0
 platform: ruby
 authors:
 - Antar Azri
 - '0x1eef'
+- Christos Maris
+- Rodrigo Serrano
 bindir: bin
 cert_chain: []
 date: 1980-01-02 00:00:00.000000000 Z
@@ -192,26 +194,65 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.7'
-description: llm.rb is a zero-dependency Ruby toolkit for Large Language Models that
-  includes OpenAI, Gemini, Anthropic, xAI (grok), zAI, DeepSeek, Ollama, and LlamaCpp.
-  The toolkit includes full support for chat, streaming, tool calling, audio, images,
-  files, and structured outputs.
+description: |
+  llm.rb is a Ruby-centric toolkit for building real LLM-powered systems — where
+  LLMs are part of your architecture, not just API calls. It gives you explicit
+  control over contexts, tools, concurrency, and providers, so you can compose
+  reliable, production-ready workflows without hidden abstractions.
+
+  Built for engineers who want to understand and control their LLM systems. No
+  frameworks, no hidden magic — just composable primitives for building real
+  applications, from scripts to full systems like Relay.
+
+  ## Key Features
+
+  - **Contexts are central** — Hold history, tools, schema, usage, cost, persistence, and execution state
+  - **Tool execution is explicit** — Run local, provider-native, and MCP tools sequentially or concurrently
+  - **One API across providers** — Unified interface for OpenAI, Anthropic, Google, xAI, zAI, DeepSeek, Ollama, and LlamaCpp
+  - **Thread-safe where it matters** — Providers are shareable, while contexts stay isolated and stateful
+  - **Production-ready** — Cost tracking, observability, persistence, and performance tuning built in
+  - **Stdlib-only by default** — Runs on Ruby standard library, with optional features loaded only when used
+
+  ## Capabilities
+
+  - Chat & Contexts with persistence
+  - Streaming responses
+  - Tool calling with JSON Schema validation
+  - Concurrent execution (threads, fibers, async tasks)
+  - Agents with auto-execution
+  - Structured outputs
+  - MCP (Model Context Protocol) support
+  - Multimodal inputs (text, images, audio, documents)
+  - Audio generation, transcription, translation
+  - Image generation and editing
+  - Files API for document processing
+  - Embeddings and vector stores
+  - Local model registry for capabilities, limits, and pricing
 email:
 - azantar@proton.me
-- 0x1eef@proton.me
+- 0x1eef@hardenedbsd.org
 executables: []
 extensions: []
 extra_rdoc_files: []
 files:
 - LICENSE
 - README.md
+- data/anthropic.json
+- data/deepseek.json
+- data/google.json
+- data/openai.json
+- data/xai.json
+- data/zai.json
 - lib/llm.rb
 - lib/llm/agent.rb
 - lib/llm/bot.rb
 - lib/llm/buffer.rb
 - lib/llm/client.rb
+- lib/llm/context.rb
+- lib/llm/context/deserializer.rb
 - lib/llm/contract.rb
 - lib/llm/contract/completion.rb
+- lib/llm/cost.rb
 - lib/llm/error.rb
 - lib/llm/eventhandler.rb
 - lib/llm/eventstream.rb
@@ -219,10 +260,21 @@ files:
 - lib/llm/eventstream/parser.rb
 - lib/llm/file.rb
 - lib/llm/function.rb
+- lib/llm/function/array.rb
+- lib/llm/function/fiber_group.rb
+- lib/llm/function/task_group.rb
+- lib/llm/function/thread_group.rb
 - lib/llm/function/tracing.rb
 - lib/llm/json_adapter.rb
+- lib/llm/mcp.rb
+- lib/llm/mcp/command.rb
+- lib/llm/mcp/error.rb
+- lib/llm/mcp/pipe.rb
+- lib/llm/mcp/rpc.rb
+- lib/llm/mcp/transport/stdio.rb
 - lib/llm/message.rb
 - lib/llm/mime.rb
+- lib/llm/model.rb
 - lib/llm/multipart.rb
 - lib/llm/multipart/enumerator_io.rb
 - lib/llm/object.rb
@@ -240,28 +292,29 @@ files:
 - lib/llm/providers/anthropic/response_adapter/completion.rb
 - lib/llm/providers/anthropic/response_adapter/enumerable.rb
 - lib/llm/providers/anthropic/response_adapter/file.rb
+- lib/llm/providers/anthropic/response_adapter/models.rb
 - lib/llm/providers/anthropic/response_adapter/web_search.rb
 - lib/llm/providers/anthropic/stream_parser.rb
 - lib/llm/providers/deepseek.rb
 - lib/llm/providers/deepseek/request_adapter.rb
 - lib/llm/providers/deepseek/request_adapter/completion.rb
-- lib/llm/providers/gemini.rb
-- lib/llm/providers/gemini/audio.rb
-- lib/llm/providers/gemini/error_handler.rb
-- lib/llm/providers/gemini/files.rb
-- lib/llm/providers/gemini/images.rb
-- lib/llm/providers/gemini/models.rb
-- lib/llm/providers/gemini/request_adapter.rb
-- lib/llm/providers/gemini/request_adapter/completion.rb
-- lib/llm/providers/gemini/response_adapter.rb
-- lib/llm/providers/gemini/response_adapter/completion.rb
-- lib/llm/providers/gemini/response_adapter/embedding.rb
-- lib/llm/providers/gemini/response_adapter/file.rb
-- lib/llm/providers/gemini/response_adapter/files.rb
-- lib/llm/providers/gemini/response_adapter/image.rb
-- lib/llm/providers/gemini/response_adapter/models.rb
-- lib/llm/providers/gemini/response_adapter/web_search.rb
-- lib/llm/providers/gemini/stream_parser.rb
+- lib/llm/providers/google.rb
+- lib/llm/providers/google/audio.rb
+- lib/llm/providers/google/error_handler.rb
+- lib/llm/providers/google/files.rb
+- lib/llm/providers/google/images.rb
+- lib/llm/providers/google/models.rb
+- lib/llm/providers/google/request_adapter.rb
+- lib/llm/providers/google/request_adapter/completion.rb
+- lib/llm/providers/google/response_adapter.rb
+- lib/llm/providers/google/response_adapter/completion.rb
+- lib/llm/providers/google/response_adapter/embedding.rb
+- lib/llm/providers/google/response_adapter/file.rb
+- lib/llm/providers/google/response_adapter/files.rb
+- lib/llm/providers/google/response_adapter/image.rb
+- lib/llm/providers/google/response_adapter/models.rb
+- lib/llm/providers/google/response_adapter/web_search.rb
+- lib/llm/providers/google/stream_parser.rb
 - lib/llm/providers/llamacpp.rb
 - lib/llm/providers/ollama.rb
 - lib/llm/providers/ollama/error_handler.rb
@@ -271,6 +324,7 @@ files:
 - lib/llm/providers/ollama/response_adapter.rb
 - lib/llm/providers/ollama/response_adapter/completion.rb
 - lib/llm/providers/ollama/response_adapter/embedding.rb
+- lib/llm/providers/ollama/response_adapter/models.rb
 - lib/llm/providers/ollama/stream_parser.rb
 - lib/llm/providers/openai.rb
 - lib/llm/providers/openai/audio.rb
@@ -290,6 +344,7 @@ files:
 - lib/llm/providers/openai/response_adapter/enumerable.rb
 - lib/llm/providers/openai/response_adapter/file.rb
 - lib/llm/providers/openai/response_adapter/image.rb
+- lib/llm/providers/openai/response_adapter/models.rb
 - lib/llm/providers/openai/response_adapter/moderations.rb
 - lib/llm/providers/openai/response_adapter/responds.rb
 - lib/llm/providers/openai/response_adapter/web_search.rb
@@ -300,23 +355,26 @@ files:
 - lib/llm/providers/xai.rb
 - lib/llm/providers/xai/images.rb
 - lib/llm/providers/zai.rb
+- lib/llm/registry.rb
 - lib/llm/response.rb
 - lib/llm/schema.rb
 - lib/llm/schema/array.rb
 - lib/llm/schema/boolean.rb
+- lib/llm/schema/enum.rb
 - lib/llm/schema/integer.rb
 - lib/llm/schema/leaf.rb
 - lib/llm/schema/null.rb
 - lib/llm/schema/number.rb
 - lib/llm/schema/object.rb
+- lib/llm/schema/parser.rb
 - lib/llm/schema/string.rb
 - lib/llm/schema/version.rb
 - lib/llm/server_tool.rb
 - lib/llm/session.rb
-- lib/llm/session/deserializer.rb
 - lib/llm/tool.rb
 - lib/llm/tool/param.rb
 - lib/llm/tracer.rb
+- lib/llm/tracer/langsmith.rb
 - lib/llm/tracer/logger.rb
 - lib/llm/tracer/null.rb
 - lib/llm/tracer/telemetry.rb
@@ -331,6 +389,7 @@ metadata:
   homepage_uri: https://github.com/llmrb/llm.rb
   source_code_uri: https://github.com/llmrb/llm.rb
   documentation_uri: https://0x1eef.github.io/x/llm.rb
+  changelog_uri: https://0x1eef.github.io/x/llm.rb/file.CHANGELOG.html
 rdoc_options: []
 require_paths:
 - lib
@@ -347,8 +406,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 rubygems_version: 3.6.9
 specification_version: 4
-summary: llm.rb is a zero-dependency Ruby toolkit for Large Language Models that includes
-  OpenAI, Gemini, Anthropic, xAI (grok), zAI, DeepSeek, Ollama, and LlamaCpp. The
-  toolkit includes full support for chat, streaming, tool calling, audio, images,
-  files, and structured outputs.
+summary: llm.rb is a Ruby-centric toolkit for building real LLM-powered systems —
+  where LLMs are part of your architecture, not just API calls. It gives you explicit
+  control over contexts, tools, concurrency, and providers, so you can compose reliable,
+  production-ready workflows without hidden abstractions.
 test_files: []

data/lib/llm/providers/gemini/response_adapter/models.rb

@@ -1,15 +0,0 @@
-# frozen_string_literal: true
-
-module LLM::Gemini::ResponseAdapter
-  module Models
-    include ::Enumerable
-    def each(&)
-      return enum_for(:each) unless block_given?
-      models.each { yield(_1) }
-    end
-
-    def models
-      body.models || []
-    end
-  end
-end