Integration: Thunderbolt

Use Thunderbolt as a cross-platform AI client for your Haystack pipelines through Hayhooks

Authors

Thunderbird

Table of Contents

• Overview
• Setup
• Usage
• License

Overview

Thunderbolt is an open-source, cross-platform AI client developed by MZLA Technologies (Thunderbird). It runs on web, iOS, Android, Mac, Linux, and Windows, and works with any OpenAI-compatible model endpoint — including self-hosted ones.

By exposing your Haystack pipeline through Hayhooks as an OpenAI-compatible endpoint, you can connect Thunderbolt to your pipeline and interact with it from any device — without building a frontend yourself.

Thunderbolt is designed for enterprise on-prem deployments but can be self-hosted locally for development and testing.

Setup

1. Expose your Haystack pipeline with Hayhooks

Install Hayhooks:

pip install hayhooks

Create a pipeline wrapper that implements run_chat_completion:

# pipelines/my_rag/pipeline_wrapper.py
from typing import Generator

from haystack import Pipeline
from haystack.components.builders import ChatPromptBuilder
from haystack.components.generators.chat import OpenAIChatGenerator
from haystack.dataclasses import ChatMessage

from hayhooks import BasePipelineWrapper, streaming_generator


class PipelineWrapper(BasePipelineWrapper):
    def setup(self) -> None:
        self.system_message = ChatMessage.from_system("You are a helpful assistant.")
        prompt_builder = ChatPromptBuilder()
        llm = OpenAIChatGenerator(model="gpt-4o-mini")

        self.pipeline = Pipeline()
        self.pipeline.add_component("prompt_builder", prompt_builder)
        self.pipeline.add_component("llm", llm)
        self.pipeline.connect("prompt_builder.prompt", "llm.messages")

    def run_chat_completion(self, model: str, messages: list[dict], body: dict) -> Generator:
        chat_messages = [self.system_message] + [
            ChatMessage.from_openai_dict_format(msg) for msg in messages
        ]
        return streaming_generator(
            pipeline=self.pipeline,
            pipeline_run_args={"prompt_builder": {"template": chat_messages}},
        )

Start Hayhooks:

hayhooks run --pipelines-dir ./pipelines

This exposes your pipeline at http://localhost:1416/v1 as an OpenAI-compatible endpoint. See Hayhooks OpenAI compatibility docs for details.

2. Deploy Thunderbolt

Follow the Thunderbolt deployment guide to self-host Thunderbolt with Docker Compose or Kubernetes, or run it locally for development. See the development guide to get started quickly.

Usage

Once Hayhooks is running and Thunderbolt is deployed:

1. Open Thunderbolt and go to Settings → Model Providers.
2. Add a new provider with a custom OpenAI-compatible base URL pointing to your Hayhooks server (e.g. http://localhost:1416/v1).
3. Select your Haystack pipeline as the model.
4. Start chatting — your messages are routed through Hayhooks to your Haystack pipeline.

This gives you a polished, cross-platform chat interface backed by whatever Haystack pipeline you choose — RAG, agents, or a custom workflow.

License

Thunderbolt is licensed under the Mozilla Public License 2.0. Hayhooks is licensed under the Apache-2.0 license.