> **Building with AI coding agents?** Install the authstack plugin with one command. This equips your agent with accurate Scalekit implementation patterns.
>
> **Recommended** (the `-y` flags skip prompts, so the command runs without an interactive terminal):
> ```bash
> npx -y @scalekit-inc/cli setup -y
> ```
>
> Global:
> ```bash
> npm install -g @scalekit-inc/cli
> scalekit setup -y
> ```
>
> Supports Claude Code, Cursor, GitHub Copilot, Codex + skills for other Agent Skills-compatible agents in `~/.agents/skills`.
> Skills: integrate-agentkit, implement-saaskit, add-mcp-oauth, implement-sso, implement-scim.
> [Full setup guide](https://docs.scalekit.com/dev-kit/build-with-ai/)

---

# Connect AI agents to GroqCloud

The GroqCloud connector lets your AI agent act in each user's GroqCloud account. Each user connects their own GroqCloud access token once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 12 tools.

**Authentication:** Bearer Token
**Categories:** AI, Developer Tools
**Tools:** 12: 7 read, 3 write, 2 destructive
**Users sign in with:** Access token
**Built by:** Scalekit connector
**Try it:** [Playground in the Scalekit dashboard](https://app.scalekit.com/ws/signup?sk_intent=playground&provider=GROQ)

## Setup

1. ### Install the SDK

   **Node.js**

   ```bash
   npm install @scalekit-sdk/node dotenv
   ```

   **Python**

   ```bash
   pip install scalekit-sdk-python python-dotenv
   ```

2. ### Set your credentials

   Add your Scalekit credentials to your `.env` file. Find values in **[app.scalekit.com](https://app.scalekit.com)** > **Developers** > **API Credentials**.

   ```sh title=".env"
   SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
   SCALEKIT_CLIENT_ID=<your-client-id>
   SCALEKIT_CLIENT_SECRET=<your-client-secret>
   ```

3. ### Create the GroqCloud connection

   In **AgentKit > Connections**, create a GroqCloud connection. The name you give it is the `connection_name` your code passes. See [Configure connections](/agentkit/connections/).

4. ### Authorize a user and make your first call

   **Node.js** (`quickstart.mts`)

   ```typescript
   import { ScalekitClient } from '@scalekit-sdk/node'
   import 'dotenv/config'
   import { createInterface } from 'node:readline/promises'

   const scalekit = new ScalekitClient(
     process.env.SCALEKIT_ENVIRONMENT_URL,
     process.env.SCALEKIT_CLIENT_ID,
     process.env.SCALEKIT_CLIENT_SECRET,
   )
   const actions = scalekit.actions

   const connector = 'groq'
   const identifier = 'user_123'

   // Generate an authorization link for the user
   const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
   console.log('Authorize GroqCloud:', link)
   const rl = createInterface({ input: process.stdin, output: process.stdout })
   await rl.question('Press Enter after authorizing...')
   rl.close()

   // Make your first call
   const result = await actions.executeTool({
     connector,
     identifier,
     toolName: 'groq_batches_list',
     toolInput: {},
   })
   console.log(result)
   ```

   ```bash
   npx tsx quickstart.mts
   ```

   **Python** (`quickstart.py`)

   ```python
   import os
   from scalekit import ScalekitClient
   from dotenv import load_dotenv
   load_dotenv()

   scalekit_client = ScalekitClient(
       env_url=os.getenv("SCALEKIT_ENVIRONMENT_URL"),
       client_id=os.getenv("SCALEKIT_CLIENT_ID"),
       client_secret=os.getenv("SCALEKIT_CLIENT_SECRET"),
   )
   actions = scalekit_client.actions

   connection_name = "groq"
   identifier = "user_123"

   # Generate an authorization link for the user
   link_response = actions.get_authorization_link(
       connection_name=connection_name,
       identifier=identifier,
   )
   print("Authorize GroqCloud:", link_response.link)
   input("Press Enter after authorizing...")

   # Make your first call
   result = actions.execute_tool(
       tool_input={},
       tool_name="groq_batches_list",
       connection_name=connection_name,
       identifier=identifier,
   )
   print(result)
   ```

   ```bash
   python quickstart.py
   ```

   Each user opens the link once and enters their GroqCloud credentials there. See [Authorize a user](/agentkit/tools/authorize/) for the full flow and statuses.

## Tools

Pass the exact name to `execute_tool`, with an input like each tool's example. To ask for a tool that's missing, use the [request form](https://scalekitsupport.portal.usepylon.com/forms/request-a-connector-tool).

**Node.js**

```typescript
const result = await actions.executeTool({
  toolName: 'groq_batch_get',
  toolInput: {
    batch_id: '<batch_id>',
  },
  connector: 'groq',
  identifier: 'user_123',
})
```

**Python**

```python
result = actions.execute_tool(
    tool_name="groq_batch_get",
    tool_input={
        "batch_id": "<batch_id>",
    },
    connection_name="groq",
    identifier="user_123",
)
```

`result.data` is the app's response as JSON, and `result.execution_id` (`executionId` in Node.js) is the ID of the call.

### `groq_batch_get`

Get Batch · Read-only

Get a batch job's status and progress.
Returns id, status, input_file_id, output_file_id, error_file_id, request_counts (total, completed, failed) and timestamps.
Use groq_file_content_get on output_file_id to read results once status is completed.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `batch_id` (`string`, required): ID of the batch. Example: batch_01jh6xa7reempvjyh6n3yst2zw.

Example input: `{"batch_id":"<batch_id>"}`

### `groq_batches_list`

List Batches · Read-only

List your organization's batch jobs.
Returns data with each batch's id, status, input_file_id, output_file_id, request_counts and timestamps.
Use groq_batch_get for one batch.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `groq_file_content_get`

Get File Content · Read-only

Download the contents of a file, such as a batch's JSONL output or error file.
Returns the raw file content (JSON Lines for batch files).
Use this after groq_batch_get shows an output_file_id.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `file_id` (`string`, required): ID of the file to download. Example: file_01jh6x76wtemjr74t1fh0faj5t.

Example input: `{"file_id":"<file_id>"}`

### `groq_file_get`

Get File · Read-only

Get one file's metadata.
Returns id, object, bytes, created_at, filename and purpose.
Use groq_file_content_get for its contents.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `file_id` (`string`, required): ID of the file. Example: file_01jh6x76wtemjr74t1fh0faj5t.

Example input: `{"file_id":"<file_id>"}`

### `groq_files_list`

List Files · Read-only

List the files uploaded to your GroqCloud organization, such as batch input and output files.
Returns data with each file's id, filename, purpose, bytes and created_at.
Use groq_file_get for one file and groq_file_content_get to read it.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `groq_model_get`

Get Model · Read-only

Get details of one GroqCloud model.
Returns id, object, created, owned_by, active, context_window and max_completion_tokens.
Use groq_models_list to see every model.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `model` (`string`, required): ID of the model. Example: llama-3.3-70b-versatile.

Example input: `{"model":"<model>"}`

### `groq_models_list`

List Models · Read-only

List the models available to your GroqCloud organization.
Returns data with each model's id, owned_by, active status, context_window and max_completion_tokens.
Use this to pick a model ID for groq_chat_completion_create. Use groq_model_get for one model.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `groq_batch_create`

Create Batch · Write

Create a batch job that runs many chat completion requests from an uploaded JSONL file asynchronously, at a discount.
Returns the batch with id, status (validating at first), input_file_id, completion_window, created_at and request_counts.
The input file must already be uploaded with purpose batch (in the GroqCloud console or API). Track it with groq_batch_get.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `completion_window` (`string`, required): Time frame for processing, from 24h to 7d. Example: 24h.
- `input_file_id` (`string`, required): ID of the uploaded JSONL file of requests, from groq_files_list. Example: file_01jh6x76wtemjr74t1fh0faj5t.
- `metadata` (`object`, optional): Custom metadata for the batch. Example: {"job": "nightly-eval"}

Example input: `{"completion_window":"<completion_window>","input_file_id":"<input_file_id>"}`

### `groq_chat_completion_create`

Create Chat Completion · Write

Generate a model reply to a conversation on GroqCloud's fast inference, with optional tool calling, JSON or structured output, reasoning and web search for compound models.
Returns id, model, choices (each with message content, tool_calls and finish_reason), usage token counts and timing information.
Use this for chat-style prompts and tool use. Use groq_response_create for the OpenAI Responses API shape.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `messages` (`array`, required): Conversation so far, oldest first. Each message has a role (system, user, assistant or tool) and content; tool results also need tool_call_id. Example: [{"role": "system", "content": "You are a concise assistant."}, {"role": "user", "content": "Summarize the benefits of LPUs in two sentences."}]
- `model` (`string`, required): ID of the model to use. Get IDs from groq_models_list or https://console.groq.com/docs/models. Example: llama-3.3-70b-versatile.
- `citation_options` (`string`, optional): Whether the model cites documents or search results. Default enabled. Example: enabled. One of: `enabled`, `disabled`.
- `compound_custom` (`object`, optional): Custom model and tool configuration for Groq Compound systems. Example: {"tools": {"enabled_tools": ["web_search"]}}
- `documents` (`array`, optional): Documents the model can cite, each an object with text (and optional id). Example: [{"id": "doc1", "text": "GroqCloud runs open models on LPUs."}]
- `frequency_penalty` (`number`, optional): Penalty from -2 to 2 for tokens by how often they already appeared. Default 0. Example: 0.5.
- `include_reasoning` (`boolean`, optional): Include a reasoning field in the reply for reasoning models. Example: false.
- `max_completion_tokens` (`integer`, optional): Maximum number of tokens to generate. Example: 1024.
- `parallel_tool_calls` (`boolean`, optional): Allow several tool calls in one turn. Default true. Example: true.
- `presence_penalty` (`number`, optional): Penalty from -2 to 2 for tokens that already appeared. Default 0. Example: 0.5.
- `reasoning_effort` (`string`, optional): Reasoning effort for reasoning models; the values each model accepts differ. Example: low. One of: `none`, `default`, `minimal`, `low`, `medium`, `high`, `xhigh`, `max`.
- `reasoning_format` (`string`, optional): How reasoning appears in the reply: hidden, raw (inline in content) or parsed (separate reasoning field). Do not combine with include_reasoning. Example: parsed. One of: `hidden`, `raw`, `parsed`.
- `response_format` (`object`, optional): Output format. {"type": "json_object"} for JSON mode, or {"type": "json_schema", "json_schema": {...}} for structured output on supported models. Example: {"type": "json_object"}
- `search_settings` (`object`, optional): Web search settings for compound models, such as include_domains, exclude_domains and country. Example: {"include_domains": ["groq.com"], "country": "united states"}
- `seed` (`integer`, optional): Seed for best-effort deterministic sampling. Example: 42.
- `service_tier` (`string`, optional): Processing tier: auto, on_demand, flex or performance. Example: auto. One of: `auto`, `on_demand`, `flex`, `performance`.
- `stop` (`array`, optional): Up to 4 sequences that stop generation. Example: ["\n\n", "END"]
- `temperature` (`number`, optional): Sampling temperature from 0 to 2. Default 1; lower is more focused. Example: 0.2.
- `tool_choice` (`string`, optional): Whether the model calls a tool: none, auto (default when tools are given) or required. Example: auto. One of: `none`, `auto`, `required`.
- `tools` (`array`, optional): Functions the model may call, each {"type": "function", "function": {name, description, parameters JSON Schema}}. Example: [{"type": "function", "function": {"name": "get_weather", "description": "Get weather for a city", "parameters": {"type": "object", "properties": {"city": {"...
- `top_p` (`number`, optional): Nucleus sampling probability mass from 0 to 1. Default 1. Change this or temperature, not both. Example: 0.9.
- `user` (`string`, optional): Your identifier for the end user, for abuse monitoring. Example: user-1234.

Example input: `{"messages":[{"role":"system","content":"You are a concise assistant."},{"role":"user","content":"Summarize the benefits of LPUs in two sentences."}],"model":"<model>"}`

### `groq_response_create`

Create Response · Write

Generate a model response with the OpenAI-compatible Responses API (beta) on GroqCloud.
Returns id, model, status, output items (messages, reasoning and function calls), output_text and usage.
Use this when your code expects the Responses API shape. Use groq_chat_completion_create for chat messages. Responses are not stored.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `input` (`string`, required): Text input to the model. Example: Write a haiku about fast inference.
- `model` (`string`, required): ID of the model to use. Get IDs from groq_models_list or https://console.groq.com/docs/models. Example: llama-3.3-70b-versatile.
- `instructions` (`string`, optional): System or developer message inserted first in the model's context. Example: Answer in one short paragraph.
- `max_output_tokens` (`integer`, optional): Upper bound on generated tokens, including reasoning tokens. Example: 512.
- `metadata` (`object`, optional): Up to 16 custom key-value pairs. Example: {"ticket": "T-1001"}
- `parallel_tool_calls` (`boolean`, optional): Allow several tool calls at once. Default true. Example: true.
- `reasoning` (`object`, optional): Reasoning configuration for reasoning models, for example {"effort": "low"}. Example: {"effort": "low"}
- `service_tier` (`string`, optional): Processing tier: auto, default or flex. Default auto. Example: auto. One of: `auto`, `default`, `flex`.
- `temperature` (`number`, optional): Sampling temperature from 0 to 2. Default 1. Example: 0.7.
- `text` (`object`, optional): Output format: plain text or structured JSON, as {"format": {"type": "text"}} or a json_schema format. Example: {"format": {"type": "text"}}
- `tool_choice` (`string`, optional): Whether the model calls a tool: none, auto or required. Example: auto. One of: `none`, `auto`, `required`.
- `tools` (`array`, optional): Function tools the model may call, at most 128, each {"type": "function", "name", "description", "parameters"}. Example: [{"type": "function", "name": "get_weather", "description": "Get weather for a city", "parameters": {"type": "object", "properties": {"city": {"type": "strin...
- `top_p` (`number`, optional): Nucleus sampling probability mass from 0 to 1. Default 1. Example: 0.9.
- `truncation` (`string`, optional): auto drops older context when the input is too long; disabled (default) returns an error. Example: disabled. One of: `auto`, `disabled`.
- `user` (`string`, optional): Your identifier for the end user. Example: user-1234.

Example input: `{"input":"<input>","model":"<model>"}`

### `groq_batch_cancel`

Cancel Batch · Destructive

Cancel a batch job that is still in progress.
Returns the batch with status cancelling, then cancelled.
Requests that already finished stay available in the output file.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `batch_id` (`string`, required): ID of the batch to cancel. Example: batch_01jh6xa7reempvjyh6n3yst2zw.

Example input: `{"batch_id":"<batch_id>"}`

### `groq_file_delete`

Delete File · Destructive

Permanently delete a file from your GroqCloud organization.
Returns id, object and deleted true.
Use groq_files_list to find file IDs.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `file_id` (`string`, required): ID of the file to delete. Example: file_01jh6x76wtemjr74t1fh0faj5t.

Example input: `{"file_id":"<file_id>"}`

## Related

Other AI connectors ([all 216](/agentkit/connectors/?category=ai)):

- [Chorus](/agentkit/connectors/chorus/): Scalekit connector, Username and password, 9 tools
- [Gong](/agentkit/connectors/gong/): Scalekit connector, OAuth, 69 tools
- [Jiminny](/agentkit/connectors/jiminny/): Scalekit connector, Access token, 27 tools


---

## More Scalekit documentation

| Resource | What it contains | When to use it |
|----------|-----------------|----------------|
| [/llms.txt](/llms.txt) | Structured index with routing hints per product area | Start here — find which documentation set covers your topic before loading full content |
| [/llms-full.txt](/llms-full.txt) | Complete documentation for all Scalekit products in one file | Use when you need exhaustive context across multiple products or when the topic spans several areas |
| [sitemap-0.xml](https://docs.scalekit.com/sitemap-0.xml) | Full URL list of every documentation page | Use to discover specific page URLs you can fetch for targeted, page-level answers |
