OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

LLM

LLM is an open-source runtime for building agentic AI applications on CRuby. It provides a unified API spanning 14+ model providers and supports concurrent and parallel tool execution, with zero runtime dependencies by default.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

LLM — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use LLM when building an agent-backed application in Ruby and you need a managed tool loop, a common API across providers, or concurrent tool execution.

Who it is for

  • Ruby developers building agentic AI applications.
  • Teams that want a shared interface across multiple model providers.
  • Developers who want agents to invoke tools implemented in their own Ruby code.

Documented uses

  • Build an agent that uses a selected model provider through the runtime's common API.
  • Define tools that let an agent interact with application capabilities such as files or other developer-written code.
  • Run independent tools concurrently or in parallel.
  • Stream agent output and receive callbacks for events such as tool calls, returns, retries and turn steps.
Before you choose

Requirements & limits

The repository specifies an MIT license. Installation requires Ruby 3.4 or later and uses the RubyGems command `gem install llm.rb`.

  • The runtime targets CRuby and requires Ruby 3.4 or later.
  • The quick-start example supplies a provider key through an environment variable.
  • Tools are implemented by developers as Ruby classes; the model decides when and how to call them.
  • A simple IO stream receives content, while a custom LLM::Stream subclass can receive additional runtime callbacks.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install llm.rb and start a basic agent conversation

Install the Ruby runtime and use its agent interface to send a prompt to DeepSeek.

Before you start

  • CRuby with Ruby 3.4 or later
  • A DeepSeek API key available in the KEY environment variable

Steps

  1. Install the gem with gem install llm.rb.

  2. In a Ruby file, require the library with require "llm".

  3. Create a provider client with LLM.deepseek(key: ENV["KEY"]) and an agent with LLM::Agent.new(llm, stream: $stdout).

  4. Call agent.talk "hello world" to start a conversation.

Expected outcome

A basic agent conversation is initiated, with streamed output directed to standard output.

Guide limitations

  • This example uses DeepSeek; the source does not document configuration for other providers here.
  • The source recommends LLM::Agent as the default high-level interface and does not establish provider access or successful execution.
Documentation 1 ↗Documentation 2 ↗
Source-based practical guide

Add a file-reading tool to an agent

Define a tool that reads a requested file and make it available to an agent.

Before you start

  • Ruby 3.4 or later with llm.rb installed
  • A DeepSeek API key available in the KEY environment variable

Steps

  1. Define a class inheriting from LLM::Tool with the name read-file, a description, and a required string path parameter.

  2. Implement call(path:) to return the file contents, as in {contents: File.read(path)}.

  3. Create a DeepSeek client and initialize LLM::Agent with tools: [ReadFile] and stream: $stdout.

  4. Ask the agent to summarize a file, for example with agent.talk "summarize README.md".

Expected outcome

The agent is configured with a file-reading tool that the model can call to access file contents.

Guide limitations

  • The tool reads files using the Ruby process's filesystem access; the source does not describe access controls or sandboxing.
  • The model decides when and how to call the tool, so the example does not guarantee that a tool call occurs.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
r-uby-dev/llm · Official homepage title · retained in its source language
How we review sourcesSuggest a correction