[PAID] 🤖 Groq Extension- Interact with multiple LLMs with Agentic Capabilities and MCP Servers- LLMs, TTS & STT all in one extension

Extensions
, , , ,
1

👨‍💻 Groq Extension

Integrate LLMs like Groq Compound, Qwen, Llama, Gemma, DeepSeek, GPT OSS & more with free limits daily! Supports adding unlimited MCP Servers, Tools like Browser Search, Browser Automation, Code Execution, Site Visiting, Wolfram Alpha & more.

:memo: Specifications

:package: Package: in.sarthakdev.groq
:gear: Version: 1.0
:iphone: Minimum API Level: 14
:date: Updated On: 2025-09-27T18:30:00Z
:computer: Built & documented using: FAST v5.0.0

:feather: Features of the Extension

  • Integrate 15+ LLMs with generous free limits daily.
  • Support for adding remote MCP servers- Tools that help you to automate tasks by connecting your LLM to other real world applications and letting it control them.
  • In built Tools for web search, code execution etc.
  • Chat History maintainance.
  • Real time Streaming of outputs.
  • JSON mode for responses in JSON.

:handshake: Multi-Components: For making use easy!

Every LLM has specific features, some have Tool use, some have MCP Servers, for some Both! So to simplify integration of LLMs into your app. I have split every LLM Type into a stable, working component. This means when you import the extension, you will get 12 Components with some even having variants, so you get all leading models from Groq.

  1. Compound (Groq): Agentic model from Groq that uses the best open source LLMs under the hood!
  2. Compound Mini (Groq): Agentic model from Groq that uses the best open source LLMs under the hood!
  3. DeepSeek (Meta + DeepSeek): The famous OS LLM now distilled with LLama.
  4. Gemma (Google): The premier OS LLM from Google.
  5. GPT OSS 120B (OpenAI): The first set of OSS models from OpenAI.
  6. GptOss20B (OpenAI): The first set of OSS models from OpenAI.
  7. Kimi K2 (Kimi): The latest trending model from Kimi.
  8. Llama 3 (Meta): The great LLMs from Meta. Known for speed and performance. (Available in two variants)
  9. Llama 4 (Meta): The great LLMs from Meta. Known for speed and performance. (Available in two variants)
  10. Llama Guard (Meta): Great models for prompt blocking and great LLMs from Meta. Known for speed and performance.
  11. Llama Prompt Guard (Meta): Great models for prompt blocking and great LLMs from Meta. Known for speed and performance. (Available in two variants)
  12. Qwen (Alibaba): A great model from Alibaba.

Feature Comparison :star2:

LLM MCP Tool Use JSON Mode Streaming Responses
GPT OSS 120B :white_check_mark: Code Interpreter, Browser Search :white_check_mark: :white_check_mark:
GPT OSS 120B :white_check_mark: Code Interpreter, Browser Search :white_check_mark: :white_check_mark:
Qwen :white_check_mark: ---- :white_check_mark: :white_check_mark:
Llama 4 :white_check_mark: ---- :white_check_mark: :white_check_mark:
Kimi K2 :white_check_mark: ---- :white_check_mark: :white_check_mark:
Compound ---- Web Search, Visit website, Browser Automation, Code Execution. ---- :white_check_mark:
Compound Mini ---- Web Search, Visit website, Browser Automation, Code Execution. --- :white_check_mark:
DeepSeek ---- ----- :white_check_mark: :white_check_mark:
Gemma ---- ---- :white_check_mark: :white_check_mark:
Llama 3 ---- ---- :white_check_mark: :white_check_mark:
Llama Guard ---- ---- :white_check_mark: :white_check_mark:
Llama Prompt Guard ---- ---- ---- ----

General Blocks :saxophone:

These blocks are present in every component, Generally. (Depends on Feature capability)

Block Details for GPT OSS


Events:

GptOss120B has total 5 events.

1. GotResponse

Fired when the complete AI response is ready. Returns the full response text, usage statistics, response time in milliseconds, and model name.

Parameter Type
response text
usage dictionary
latencyMs number
model text

2. GotStream

Fired during streaming with each text chunk. Use 'chunk' to append to your label for real-time display. When 'done' is true, streaming has finished.

Parameter Type
chunk text
index number
done boolean

3. ErrorOccurred

Fired when an error occurs. Returns error code, human-readable message, and raw error response for debugging.

Parameter Type
code text
message text
raw text

4. GotJSON

Fired when JSON mode is enabled and response is successfully parsed. Returns structured data as dictionary and raw JSON string.

Parameter Type
json dictionary
raw text

5. GotReasoning

Fired when the AI's reasoning/thinking process is available. Shows the model's step-by-step thought process before generating the response.

Parameter Type
reasoning text

Methods:

GptOss120B has total 9 methods.

1. AddMcpServer

Add an MCP server tool with label, url and headers (YailDictionary).

Parameter Type
serverLabel text
serverUrl text
headers dictionary

2. ClearMcpServers

Remove all MCP servers (tools list).

3. GetMcpServers

Get MCP servers as list of dictionaries.

  • Return type: list

4. Ask

Send a message to the AI and get a response. The prompt is your question or instruction to the AI.

Parameter Type
prompt text

5. AskWithSystem

Send a message with a system instruction. System sets the AI's behavior/role, prompt is your question.

Parameter Type
system text
prompt text

6. AddMessage

Append a message to the internal history. Role must be user/assistant/system.

Parameter Type
role text
content text

7. ClearConversation

Clear all conversation history. Use this to start a fresh conversation with the AI.

8. GetStreamingText

Get the accumulated streaming text so far. Useful for displaying the full response as it builds up.

  • Return type: text

9. Cancel

Stop the current AI request immediately. Useful for stopping long responses or streaming.

Designer:

GptOss120B has total 20 designer properties.

1. ApiKey

  • Input type: string

2. BaseUrl

  • Input type: string
  • Default value: https://api.groq.com/openai/v1

3. Temperature

  • Input type: float
  • Default value: 1

4. IncludeTemperature

  • Input type: boolean
  • Default value: True

5. TopP

  • Input type: float
  • Default value: 1

6. IncludeTopP

  • Input type: boolean
  • Default value: True

7. MaxTokens

  • Input type: non_negative_integer
  • Default value: 8192

8. IncludeMaxTokens

  • Input type: boolean
  • Default value: True

9. Stream

  • Input type: boolean
  • Default value: False

10. IncludeStream

  • Input type: boolean
  • Default value: True

11. HistoryEnabled

  • Input type: boolean
  • Default value: False

12. HistoryLimit

  • Input type: non_negative_integer
  • Default value: 20

13. JSONMode

  • Input type: boolean
  • Default value: False

14. ReasoningLevel

  • Input type: string
  • Default value: medium

15. IncludeReasoning

  • Input type: boolean
  • Default value: True

16. IncludeResponseFormat

  • Input type: boolean
  • Default value: True

17. IncludeStopNull

  • Input type: boolean
  • Default value: False

18. IncludeTools

  • Input type: boolean
  • Default value: True

19. ToolCodeInterpreter

  • Input type: boolean
  • Default value: True

20. ToolBrowserSearch

  • Input type: boolean
  • Default value: True

Setters:

GptOss120B has total 20 setter properties.

1. ApiKey

Your Groq API key.

  • Input type: text

2. BaseUrl

Base URL (OpenAI-compatible). Default: https://api.groq.com/openai/v1

  • Input type: text

3. Temperature

Temperature 0..2 (default 1)

  • Input type: number

4. IncludeTemperature

Include temperature field.

  • Input type: boolean

5. TopP

TopP 0..1 (default 1)

  • Input type: number

6. IncludeTopP

Include top_p field.

  • Input type: boolean

7. MaxTokens

Maximum completion tokens (default 8192)

  • Input type: number

8. IncludeMaxTokens

Include max_completion_tokens field.

  • Input type: boolean

9. Stream

Enable streaming (disabled when JSONMode is true).

  • Input type: boolean

10. IncludeStream

Include stream field.

  • Input type: boolean

11. HistoryEnabled

Maintain conversation history (default False).

  • Input type: boolean

12. HistoryLimit

History turn limit (default 20). Each turn adds user and assistant messages.

  • Input type: number

13. JSONMode

JSON mode: when true, response_format=json_object and streaming is disabled.

  • Input type: boolean

14. ReasoningLevel

Reasoning level: low, medium, or high.

  • Input type: text

15. IncludeReasoning

Include reasoning_effort field.

  • Input type: boolean

16. IncludeResponseFormat

Include response_format when JSONMode is true.

  • Input type: boolean

17. IncludeStopNull

Include stop: null (send explicit JSON null).

  • Input type: boolean

18. IncludeTools

Include tools array.

  • Input type: boolean

19. ToolCodeInterpreter

Include code_interpreter tool.

  • Input type: boolean

20. ToolBrowserSearch

Include browser_search tool.

  • Input type: boolean

Getters:

GptOss120B has total 20 getter properties.

1. ApiKey

Your Groq API key.

  • Return type: text

2. BaseUrl

Base URL (OpenAI-compatible). Default: https://api.groq.com/openai/v1

  • Return type: text

3. Temperature

Temperature 0..2 (default 1)

  • Return type: number

4. IncludeTemperature

Include temperature field.

  • Return type: boolean

5. TopP

TopP 0..1 (default 1)

  • Return type: number

6. IncludeTopP

Include top_p field.

  • Return type: boolean

7. MaxTokens

Maximum completion tokens (default 8192)

  • Return type: number

8. IncludeMaxTokens

Include max_completion_tokens field.

  • Return type: boolean

9. Stream

Enable streaming (disabled when JSONMode is true).

  • Return type: boolean

10. IncludeStream

Include stream field.

  • Return type: boolean

11. HistoryEnabled

Maintain conversation history (default False).

  • Return type: boolean

12. HistoryLimit

History turn limit (default 20). Each turn adds user and assistant messages.

  • Return type: number

13. JSONMode

JSON mode: when true, response_format=json_object and streaming is disabled.

  • Return type: boolean

14. ReasoningLevel

Reasoning level: low, medium, or high.

  • Return type: text

15. IncludeReasoning

Include reasoning_effort field.

  • Return type: boolean

16. IncludeResponseFormat

Include response_format when JSONMode is true.

  • Return type: boolean

17. IncludeStopNull

Include stop: null (send explicit JSON null).

  • Return type: boolean

18. IncludeTools

Include tools array.

  • Return type: boolean

19. ToolCodeInterpreter

Include code_interpreter tool.

  • Return type: boolean

20. ToolBrowserSearch

Include browser_search tool.

  • Return type: boolean

Get the extension now! :grinning:

The extension is available for purchase at just 7.99$ instantly!

If you have any questions, feel free to ask below!

2

:crystal_ball: Future of PlayAI and GroqText Extension!

GroqText extension (Older one) is now deprecated and replaced by this modern Groq Extension- Highly Improved with extensive features. In a similar manner, PlayAI extension will also be deprecated in the near future. But don't worry all its features will be integrated in this AI Extension** and it will be available to the existing users of GroqText and PlayAI Extension for Free of Cost!

This means, this extension will get all the features of -> GroqText (5.99$), GroqVision (5.99$) and PlayAI (7.99$) extensions for only one third of their total cost :star_struck:

Spoiler Alert: In the next updates, this extension will also get a new OpenAI Whisper component for Speech to Text transcription using the best model from OpenAI! It may be possible that prices go high then, so what are you waiting for buy the extension now!

3

All About MCP Servers!

An MCP server is a server that implements the Model Context Protocol (MCP), an open-source standard for connecting AI agents with external tools and data. It exposes the capabilities and resources of underlying systems, allowing AI agents to interact with services like databases, APIs, and development tools without needing to understand the complex code and authentication involved. By abstracting these complexities, MCP servers simplify the integration of external data and functions for AI applications, enhancing their ability to perform complex tasks.

You can use the Groq extension with an MCP server to enable extremely fast AI agents that interact with external tools and data sources. By combining Groq's high-speed inference engine with the standardized Model Context Protocol, developers can build agents that perform complex, multi-step tasks with minimal latency.

Example use cases with Groq Extension MCP

  1. Multi-agent collaboration with AutoGen
  • Prompt: A developer asks for help generating a new project.
  • Action: A multi-agent framework like AutoGen, powered by Groq's speed, orchestrates a team of AI agents. One agent may use a GitHub MCP server to create a new project, another may use a Code Generation MCP server to write boilerplate code, and a third may use a Search MCP server to find relevant libraries.
  • Result: A fully formed, initialized project is ready almost instantly, complete with generated code and project infrastructure.
  1. Lead qualification and outreach: An AI agent can automate the process of finding and engaging with potential clients, dramatically reducing the manual workload for sales teams.
  • Groq Agent: Qualify new leads by integrating with a spreadsheet or CRM and a web-scraping MCP server.
  • Prompt: "For all new leads in the 'New Leads' spreadsheet, research their company on the web to find relevant recent news. For leads in the tech industry, draft a personalized outreach email mentioning a recent company announcement. Send the drafted emails via the Gmail MCP server and move the leads to the 'Engaged' sheet."
  • Behind the scenes:
    1. The agent uses a Google Sheets MCP server to read new leads.
    2. For each lead, it uses a Firecrawl or Browserbase MCP server to browse the web for news.
    3. It then uses a Gmail MCP server to send the email and a Google Sheets MCP server to update the lead's status.
4

Sorry, I think I posted it in the wrong category. Posted after 4-5 months almost. Please take it to the extensions category. Thanks and sorry.

5

I am very much interested to see what you build with this :blush:

6

:loudspeaker: Big Update is here!

8agaDw54_400x400
8agaDw54_400x4001080×1080 60.8 KB
gptoss
gptoss1080×1080 56.9 KB

:sparkles: New Components: PlayAI (TTS) & OpenAI Whisper (STT) !

  • Generate Real Life Text to Speech with PlayAI in 20 voices accelerated by Groq!
  • Convert your audio to text with OpenAI Whisper, available in two variants- Whisper & Whisper Turbo running on Groq Servers!
Blocks

blocks (4)
blocks (4)376×1240 42.7 KB
417×750 25.1 KB

Start using Groq Extension now!

(Existing users of PlayAI & GroqText extension -> Dm me to get the updated Groq extension at a discounted upgrade price)