A lightweight Ruby library for interacting with multiple LLM providers
All providers inherit from LLM::Provider.
They share a common interface and set of functionality between them. They can be
instantiated with an API key and an (optional) set of options via the
the singleton methods of LLM.
For example:
#!/usr/bin/env ruby
require "llm"
llm = LLM.openai("yourapikey", <options>)
llm = LLM.anthropic("yourapikey", <options>)
llm = LLM.ollama(nil, <options>)
# etc ...The
LLM::Provider#chat
method returns a
LLM::LazyConversation
object
that can maintain a "lazy" conversation where input prompts are sent to the
provider only when neccessary. Once a conversation is initiated it will maintain a
thread of messages that provide the LLM with a certain amount of extra information
that can be re-used within the conversation:
#!/usr/bin/env ruby
require "llm"
llm = LLM.openai("yourapikey")
bot = llm.chat "keep the answer concise", :system
bot.chat URI("https://upload.wikimedia.org/wikipedia/commons/b/be/Red_eyed_tree_frog_edit2.jpg")
bot.chat "What is the frog's name?"
bot.chat "What is the frog's habitat?"
bot.chat "What is the frog's diet?"
bot.messages.each do |message|
##
# At this point a single request is made to the provider
# See 'LLM::MessageQueue' for more details
print "[#{message.role}] ", message.content, "\n"
end
##
# [system] keep the answer concise
# [user] https://upload.wikimedia.org/wikipedia/commons/b/be/Red_eyed_tree_frog_edit2.jpg
# [user] What is the frog's name?
# [user] What is the frog's habitat?
# [user] What is the frog's diet?
# [assistant] The frog in the image is likely a Red-eyed Tree Frog.
#
# #### Habitat:
# - Typically found in tropical rainforests, especially in Central America.
#
# #### Diet:
# - Primarily insectivorous, feeding on insects like crickets and moths.The
LLM::Provider#chat
method returns a
LLM::Conversation
object that can maintain a conversation with a LLM provider but unlike
LLM::LazyConversation
each call to chat! / chat corresponds to a HTTP request to the provider:
#!/usr/bin/env ruby
require "llm"
llm = LLM.openai("yourapikey")
bot = llm.chat! "be a helpful assistant", :system
bot.chat "keep the answers short and sweet", :system
bot.chat "help me choose a good book"
bot.chat "books of poetry"
bot.messages.each do |message|
print "[#{message.role}] ", message.content, "\n"
end
##
# [system] be a helpful assistant
# [assistant] Of course! How can I assist you today?
# [system] keep the answers short and sweet
# [assistant] Got it! What do you need help with?
# [user] help me choose a good book
# [assistant] Sure! What genre are you interested in?
# [user] books of poetry
# [assistant] Here are a few great poetry collections:
#
# 1. **"The Sun and Her Flowers" by Rupi Kaur**
# 2. **"The Carrying" by Ada Limón**
# 3. **"Milk and Honey" by Rupi Kaur**
# 4. **"Ariel" by Sylvia Plath**
# 5. **"The Poetry of Pablo Neruda"**
#
# Happy reading!A complete API reference is available at 0x1eef.github.io/x/llm
LLM has not been published to RubyGems.org yet. Stay tuned
MIT. See LICENSE.txt for more details