λ ollama-haskell v0.4.1
Hackage GitHub
Haskell client for local LLMs

Type-safe AI engineering in Haskell

High-throughput Conduit streaming, automatic JSON Schema derivation via GHC Generics, Model Context Protocol (MCP) bridging, and STM conversation management.

Get Started Explore Tutorials
Haskell for Local AI

ollama-haskell brings production-ready type-safety, deterministic JSON outputs, composable streaming, and Model Context Protocol (MCP) tooling to locally hosted Large Language Models.

5-Line Quick Start

Interact with local models with zero boilerplate:

import Data.List.NonEmpty (NonEmpty ((:|)))
import Data.Text.IO qualified as TIO
import Ollama

main :: IO ()
main = do
  client <- defaultClient
  res    <- chat client $ chatRequest "qwen3.5:2b" (userMessage "Why is the sky blue?" :| [])
  case res of
    Left err   -> print err
    Right resp -> mapM_ (TIO.putStrLn . messageContent) (crMessage resp)

Core Capabilities

<div class="feature-card">
    <h3>First-Class Conduit Streaming</h3>
    <p>Stream tokens asynchronously with minimal memory footprint and zero buffering delays via <code>chatStream</code> and <code>generateStream</code>.</p>
</div>

<div class="feature-card">
    <h3>Generic JSON Schema Derivation</h3>
    <p>Derive strict JSON schemas directly from Haskell records via <code>GHC.Generics</code> with <code>ToSchema</code> and <code>formatFor</code>.</p>
</div>

<div class="feature-card">
    <h3>Model Context Protocol (MCP)</h3>
    <p>Seamless bidirectional bridging between Ollama tool calls and MCP servers via <code>mcp-server</code> over stdio and HTTP.</p>
</div>

<div class="feature-card">
    <h3>Thinking / Reasoning Models</h3>
    <p>Native ADT support for deep reasoning models like <code>qwen3.5</code> and <code>deepseek-r1</code> with <code>ThinkingLevel</code> controls.</p>
</div>

<div class="feature-card">
    <h3>STM Conversation Store</h3>
    <p>Transactional, thread-safe in-memory session management with <code>InMemoryStore</code> and <code>ConversationStore</code> typeclasses.</p>
</div>

Installation

Add ollama-haskell to your project dependencies in .cabal:

build-depends:
    base >= 4.17 && < 5
  , ollama-haskell >= 0.4.1.0

Or in Stack package.yaml:

dependencies:
  - ollama-haskell >= 0.4.1.0

Next Steps

<a href="/ollama-101.html" class="feature-card" style="text-decoration: none;">
    <h3>Ollama 101 Beginner Guide</h3>
    <p>Never used Ollama before? Learn how to install Ollama, download models, and run your first query in under 2 minutes.</p>
</a>
<a href="/tutorials/chat.html" class="feature-card" style="text-decoration: none;">
    <h3>Feature Tutorials</h3>
    <p>Hands-on guides covering chat streaming, tool calling, MCP servers, embeddings, and structured outputs.</p>
</a>