/ 
Langertha-0.100
/ Langertha::Engine::Ollama

NAME

Langertha::Engine::Ollama - Ollama API

VERSION

version 0.100

SYNOPSIS

use Langertha::Engine::Ollama;

my $ollama = Langertha::Engine::Ollama->new(
  url => $ENV{OLLAMA_URL},
  model => 'llama3.3',
  system_prompt => 'You are a helpful assistant',
  context_size => 2048,
  temperature => 0.5,
);

print($ollama->simple_chat('Say something nice'));

my $embedding = $ollama->embedding($content);

# Get OpenAI compatible API access to Ollama
my $ollama_openai = $ollama->openai;

# List available models
my $models = $ollama->simple_tags;

# Show running models
my $running = $ollama->simple_ps;

DESCRIPTION

This module provides access to Ollama, which runs large language models locally. Ollama supports many popular open-source models with various sizes and capabilities.

Popular Models (February 2026):

  • llama3.3 - Meta's Llama 3.3 70B with 128k context (default). Excellent general-purpose model with broad tool support.

  • llama3.2 - Meta's Llama 3.2 includes small models (1B, 3B) for efficient local inference.

  • qwen3 - Latest Qwen 3 generation with enhanced reasoning. Qwen3-30B recommended for most teams (delivers 90%+ flagship power at lower cost).

  • qwen2.5 - Qwen 2.5 family (up to 72B) with strong multilingual support and 128k context. Excellent for general tasks.

  • qwen2.5-coder - Qwen 2.5 specialized for code generation and programming tasks.

  • deepseek-coder-v2 - DeepSeek's coding-specialized model. Excellent for software development.

  • mixtral - Mistral's mixture-of-experts model (8x22B). Cost-effective performance.

  • mistral - Mistral models including latest Mistral 3 family (3B, 8B, 14B).

  • codestral - Mistral's code-specialized model.

  • mxbai-embed-large - Embedding model (default for embeddings).

Model Selection Tips:

  • For general tasks: qwen2.5-72b or llama3.3

  • For coding: deepseek-coder-v2 or qwen2.5-coder

  • For reasoning: llama3.3 or qwen3

  • For cost-effective performance: mixtral-8x22b

  • For low-resource systems: llama3.2-3b or qwen3-30b

Features:

  • Run models completely locally

  • No API key required

  • Chat completions with streaming

  • Embeddings

  • Custom models and quantization

  • OpenAI-compatible API access via openai() method

  • JSON format output support

  • Keep-alive model loading control

  • Dynamic model listing with caching

THIS API IS WORK IN PROGRESS

LISTING AVAILABLE MODELS

Fetch models from your local Ollama instance:

# Get simple list of model names
my $model_ids = $ollama->list_models;
# Returns: ['llama3.3', 'qwen2.5', ...]

# Get full model objects with metadata
my $models = $ollama->list_models(full => 1);

# Force refresh (bypass cache)
my $models = $ollama->list_models(force_refresh => 1);

# Or use the original method
my $tags = $ollama->simple_tags;

Caching: Results are cached for 1 hour. Configure TTL via models_cache_ttl or clear manually with clear_models_cache.

HOW TO INSTALL OLLAMA

https://github.com/ollama/ollama/tree/main

To pull a model:

ollama pull llama3.3
ollama pull qwen3

To list available models from Ollama library:

ollama list

SEE ALSO

SUPPORT

Issues

Please report bugs and feature requests on GitHub at https://github.com/Getty/langertha/issues.

CONTRIBUTING

Contributions are welcome! Please fork the repository and submit a pull request.

AUTHOR

Torsten Raudssus <torsten@raudssus.de> https://raudss.us/

COPYRIGHT AND LICENSE

This software is copyright (c) 2026 by Torsten Raudssus.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.