swarmauri_llm_groq 0.11.0.dev1

Provider-specific Swarmauri import package for Groq OpenAI-compatible chat, tool use, vision-language, streaming, and audio transcription workflows.

Swarmauri Groq LLM

swarmauri_llm_groq provides provider-specific Groq imports for Swarmauri applications that need fast OpenAI-compatible inference. The package exports chat, tool-calling, vision, and audio adapters while keeping their runtime implementations aligned with swarmauri_standard.

Groq's public API documentation describes an OpenAI-compatible base URL at https://api.groq.com/openai/v1, with endpoints for chat completions, audio transcriptions, and audio translations. This package maps those API surfaces into Swarmauri components:

• GroqModel for chat completions, streaming, JSON responses, and batch helper workflows.
• GroqToolModel for function/tool calling through Swarmauri Toolkit instances.
• GroqVisionModel for vision-language chat completion payloads.
• GroqAIAudio for transcription and translation with Groq-hosted Whisper models.

Why Use This Package?

• Keep Groq-specific model dependencies explicit while preserving common Swarmauri conversation and toolkit patterns.
• Use Groq's low-latency OpenAI-compatible chat completions without rewriting application code around raw HTTP payloads.
• Add tool calling, vision-language prompts, or speech-to-text workflows from one provider package.
• Swap between Groq and other Swarmauri LLM provider packages while retaining familiar predict, stream, apredict, astream, batch, and abatch workflows.

FAQ

What does swarmauri_llm_groq install?

It installs provider package entry points for GroqModel, GroqToolModel, GroqVisionModel, and GroqAIAudio.

Which Groq endpoints does it call?

The chat, tool, and vision adapters post to https://api.groq.com/openai/v1/chat/completions. The audio adapter uses https://api.groq.com/openai/v1/audio/transcriptions and https://api.groq.com/openai/v1/audio/translations.

Does it support tool calling?

Yes. GroqToolModel converts Swarmauri toolkit schemas to Groq-compatible tool definitions, sends them with chat completions, executes returned tool calls, and can perform a follow-up model call with tool results.

Does it support vision?

Yes. GroqVisionModel accepts Swarmauri message content that can include structured content lists for OpenAI-compatible vision chat payloads. The runtime also emits a deprecation warning that new projects should review the newer Swarmauri VLM imports when available.

Does audio streaming work?

No. GroqAIAudio supports file-based transcription and translation through predict, apredict, batch, and abatch. Its stream and astream methods raise NotImplementedError.

Features

• GroqModel for synchronous, asynchronous, streaming, JSON, and batch chat completion workflows.
• GroqToolModel for Groq tool use with Swarmauri Toolkit schemas.
• GroqVisionModel for OpenAI-compatible vision-language chat payloads.
• GroqAIAudio for Whisper transcription and translation tasks.
• Usage metadata support where Groq chat responses include usage data.
• Model discovery through Groq's OpenAI-compatible models endpoint for chat.
• Compatibility with Python 3.10, 3.11, 3.12, 3.13, and 3.14.

Installation

uv add swarmauri_llm_groq

pip install swarmauri_llm_groq

Prerequisites

Create a Groq API key in the Groq console and pass it to the model constructor as api_key=.... Package live tests are skipped unless provider credentials are available in the environment.

Usage

from swarmauri_llm_groq import GroqModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage

conversation = Conversation()
conversation.add_message(HumanMessage(content="Write a one-sentence release note."))

model = GroqModel(api_key="GROQ_API_KEY")
result = model.predict(conversation=conversation, max_tokens=120)

print(result.get_last().content)

Streaming Chat Completion

from swarmauri_llm_groq import GroqModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage

conversation = Conversation()
conversation.add_message(HumanMessage(content="List three build pipeline risks."))

model = GroqModel(api_key="GROQ_API_KEY")

for token in model.stream(conversation=conversation):
    print(token, end="")

Tool Calling

from swarmauri_llm_groq import GroqToolModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage
from swarmauri_standard.toolkits.Toolkit import Toolkit

conversation = Conversation()
conversation.add_message(HumanMessage(content="Use a tool if one is relevant."))

toolkit = Toolkit()
model = GroqToolModel(api_key="GROQ_API_KEY")
result = model.predict(conversation=conversation, toolkit=toolkit)

print(result.get_last().content)

Audio Transcription

from swarmauri_llm_groq import GroqAIAudio

audio = GroqAIAudio(api_key="GROQ_API_KEY")
text = audio.predict("meeting.wav", task="transcription")

print(text)

Vision-Language Prompt

from swarmauri_llm_groq import GroqVisionModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage

conversation = Conversation()
conversation.add_message(
    HumanMessage(
        content=[
            {"type": "text", "text": "Describe the image."},
            {"type": "image_url", "image_url": {"url": "https://example.com/image.png"}},
        ]
    )
)

model = GroqVisionModel(api_key="GROQ_API_KEY")
result = model.predict(conversation=conversation)

print(result.get_last().content)

Related Packages

• swarmauri_llm_anthropic
• swarmauri_llm_deepinfra
• swarmauri_llm_deepseek
• swarmauri_llm_gemini
• swarmauri_llm_mistral
• swarmauri_llm_openai
• swarmauri_llm_whisper

Foundational Swarmauri Packages

• swarmauri_core
• swarmauri_base
• swarmauri_standard
• swarmauri

Provider Documentation

• Groq API reference
• Groq OpenAI compatibility
• Groq models
• Groq tool use
• Groq speech to text

Best Practices

• Store Groq API keys in environment variables or a secrets manager.
• Confirm model availability against Groq's current model catalog before production deployment.
• Use GroqToolModel only when tools are needed; plain chat requests are simpler with GroqModel.
• Use GroqAIAudio for file-based transcription and translation, not streaming audio.
• Prefer newer Swarmauri VLM and STT imports for new projects when available.

License

Apache-2.0