### Start Workflow Examples

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/workflows/getting-started/cookbook_examples.md

After setting up a project, navigate to the project directory and run this command to register workflows and start the worker. This prepares the environment for running example executions.

```bash
cd my-workflow # or the name of the folder you created
make start-examples
```

--------------------------------

### Install Dependencies and Run Server

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/vibe-code/scaffold-a-project.md

After the project is scaffolded, install the necessary Python dependencies and start the Flask development server.

```bash
pip install -r requirements.txt
python app.py
```

--------------------------------

### Start the CLI

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/vibe/code/cli/work-with-cli.md

Run the CLI from your project root. You can also provide an initial prompt to guide the CLI's first action.

```bash
vibe
```

```bash
vibe "Refactor the main function in cli/main.py to be more modular."
```

--------------------------------

### Start Conversation Response Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/conversations/page.mdx

Example of a successful response when starting a new conversation.

```APIDOC
## POST /v1/conversations/start

### Description
Starts a new conversation.

### Method
POST

### Endpoint
/v1/conversations/start

### Response
#### Success Response (200)
- **conversation_id** (string) - The unique identifier for the conversation.
- **outputs** (array) - A list of outputs from the conversation.
  - **content** (string) - The content of an output.
- **usage** (object) - Information about the token usage for the conversation.

#### Response Example
```json
{
  "conversation_id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
  "outputs": [
    {
      "content": "Example content."
    }
  ],
  "usage": {}
}
```
```

--------------------------------

### Run Interactive CLI Setup

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/vibe/code/cli/api-keys-profiles/page.mdx

Initiate the CLI setup flow to configure API keys. This is the recommended method for first-time users and saves the key to `~/.vibe/.env`.

```bash
vibe
```

```bash
vibe --setup
```

--------------------------------

### Create Project Directory and Launch Vibe CLI

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/getting-started/quickstarts/vibe-code/scaffold-a-project/page.mdx

Start by creating a new, empty directory for your project and then navigate into it. Launch the Vibe CLI within this directory to begin scaffolding.

```bash
mkdir my-project && cd my-project
vibe
```

--------------------------------

### Get Users API Response Example (200 OK)

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/admin/users/page.mdx

This is an example of a successful response from the Get Users API endpoint. It includes lists of 'invites' and 'members', along with pagination details if provided.

```json
{
  "invites": [
    {
      "created_at": "2025-12-17T10:25:07.818693Z",
      "email": "ipsum eiusmod",
      "expired": false,
      "raw_role": "A",
      "raw_roles": [
        "A"
      ],
      "uuid": "019b2bd7-96e7-7219-8c0b-45a73da50088"
    }
  ],
  "members": [
    {
      "created_at": "2025-12-17T10:25:07.818693Z",
      "email": null,
      "name": null,
      "oid_id": null,
      "raw_role": "A",
      "raw_roles": [
        "A"
      ],
      "uuid": "019b2bd7-96e7-7219-8c0b-45a73da50088"
    }
  ],
  "page": null,
  "page_size": null,
  "total": null
}
```

--------------------------------

### Run Vibe CLI Setup

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/vibe/code/cli/api-keys-profiles.md

Initiate the interactive setup flow for the Vibe CLI to configure API keys. This is recommended for first-time use or to reconfigure.

```bash
vibe
```

```bash
vibe --setup
```

--------------------------------

### Response Example for Get Schedules

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/workflows/schedules/page.mdx

Example JSON response when successfully retrieving workflow schedules.

```json
{
  "schedules": [
    {
      "input": "Example input.",
      "paused": false,
      "schedule_id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
      "workflow_name": "support-workflow"
    }
  ]
}
```

--------------------------------

### Manual Install Vibe CLI with uv

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/vibe-code/install-cli.md

Recommended manual installation method for macOS, Linux, and Windows using `uv`.

```bash
uv tool install mistral-vibe
```

--------------------------------

### Document Annotation Prompt Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/document-processing/annotations.md

Provides an example of a system prompt to guide the document annotation process.

```APIDOC
## Document Annotation Prompt

### Description
An optional prompt that provides high-level system instructions for the annotation step, offering further context and guidance.

### Example (Python)
```python
document_annotation_prompt = """
Extract the following from the provided PDF document:
- Language (e.g., "English")
- All chapter/section titles (e.g., ["Abstract", "1 Introduction"])
- All URLs (e.g., ["https://example.com"])
Be precise and include only exact matches.
"""
```

### Example (TypeScript)
```typescript
const documentAnnotationPrompt = `
Extract the following from the provided PDF document:
- Language (e.g., "English")
- All chapter/section titles (e.g., ["Abstract", "1 Introduction"])
- All URLs (e.g., ["https://example.com"])
Be precise and include only exact matches.
`;
```

### Example (Curl)
```bash
"document_annotation_prompt": "Extract the following from the provided PDF document:\n- Language (e.g., \"English\")\n- All chapter/section titles (e.g., [\"Abstract\", \"1 Introduction\"])\n- All URLs (e.g., [\"https://example.com\"])\nBe precise and include only exact matches."
```
```

--------------------------------

### Basic Semantic Cache Setup

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/studio-api/search-toolkit/retrieval/semantic-cache/page.mdx

Demonstrates the core setup for a semantic cache using an in-memory backend. It involves creating a cache backend, initializing the semantic cache with a similarity threshold, and then wrapping an existing query engine with the cache.

```python
from mistralai.search.toolkit.retrieval.cache import (
    CachedQueryEngine,
    InMemoryCacheBackend,
    SemanticCache,
    EvictionPolicy,
)

# 1. Create a cache backend
backend = InMemoryCacheBackend(
    dim=1024,                          # Embedding dimensionality
    max_entries=500,                   # Max cached queries
    ttl_seconds=3600,                  # 1-hour expiration
    eviction_policy=EvictionPolicy.LRU,
)

# 2. Create the semantic cache
cache = SemanticCache(
    backend=backend,
    similarity_threshold=0.95,  # 95% cosine similarity = cache hit
)

# 3. Wrap your QueryEngine
cached_engine = CachedQueryEngine(
    engine=query_engine,
    cache=cache,
    embedder=embedder,  # Used to embed incoming queries
)

# 4. Use normally. Caching is transparent.
result = await cached_engine.search("What is RAG?", top_k=10)
# First call: embedding + retrieval → cached
# Second call with similar query: cache hit → instant
```

--------------------------------

### Request Body Example for Conversation Stream

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/conversations/page.mdx

Provides an example JSON payload for the request body when starting a conversation stream.

```json
{
  "inputs": "Example input."
}
```

--------------------------------

### Start Development Server

Source: https://github.com/mistralai/platform-docs-public/blob/main/CLAUDE.md

Starts the development server. It first builds cookbooks and exports raw MDX, then runs the Next.js development server.

```bash
pnpm dev
```

--------------------------------

### Get Deployment by Name

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/workflows/deployments/page.mdx

This example shows how to retrieve a deployment's details using its name via a GET request.

```APIDOC
## GET /v1/workflows/deployments/{name}

### Description
Retrieves the details of a specific deployment.

### Method
GET

### Endpoint
/v1/workflows/deployments/{name}

### Parameters
#### Path Parameters
- **name** (string) - Required - The name of the deployment to retrieve.

### Request Example
```curl
curl https://api.mistral.ai/v1/workflows/deployments/{name} \
 -X GET \
 -H 'Authorization: Bearer YOUR_APIKEY_HERE'
```

### Response
#### Success Response (200)
- **created_at** (date-time) - Required - When the deployment was first registered
- **id** (string) - Required - The unique identifier for the deployment
- **is_active** (boolean) - Required - Indicates if the deployment is currently active
- **name** (string) - Required - The name of the deployment
- **updated_at** (date-time) - Required - When the deployment was last updated
- **workers** (array) - Required - A list of workers associated with the deployment
  - **created_at** (date-time) - Required - When the worker was first registered
  - **is_active** (boolean) - Required - Indicates if the worker is currently active
  - **name** (string) - Required - The name of the worker
  - **updated_at** (date-time) - Required - When the worker was last updated

#### Response Example
```json
{
  "created_at": "2025-12-17T10:25:07.818693Z",
  "id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
  "is_active": false,
  "name": "My resource",
  "updated_at": "2025-12-17T10:41:03.469341Z",
  "workers": [
    {
      "created_at": "2025-12-17T10:25:07.818693Z",
      "is_active": false,
      "name": "My resource",
      "updated_at": "2025-12-17T10:41:03.469341Z"
    }
  ]
}
```
```

--------------------------------

### Launch Vibe CLI Setup Wizard

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/vibe-code/install-cli.md

Run this command to launch the Vibe CLI setup wizard, which helps in registering your API key.

```bash
vibe
```

--------------------------------

### Get Extracted Text Signed URL Response Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/libraries/documents/page.mdx

This is an example of a successful JSON response containing a signed URL.

```json
"https://example.com/signed-url"
```

--------------------------------

### Install All Search Toolkit Extras

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/studio-api/search-toolkit/page.mdx

Install the core package with all available optional extras for comprehensive functionality. This command uses uv for dependency management.

```bash
uv add "mistralai-search-toolkit[all]"
```

--------------------------------

### Install pnpm and Node on macOS

Source: https://github.com/mistralai/platform-docs-public/blob/main/README.md

Install Node.js and pnpm package manager on macOS using Homebrew. These are prerequisites for project setup.

```bash
brew install pnpm
brew install node
```

--------------------------------

### Create Directory and Launch Vibe CLI

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/vibe-code/scaffold-a-project.md

Start in a fresh directory to ensure the CLI generates the project from scratch. If you run `vibe` in an existing project with a `.vibe/` directory, the CLI will ask for confirmation.

```bash
mkdir my-project && cd my-project
vibe
```

--------------------------------

### Start Streaming Conversation with cURL

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/studio-api/agents/agents-api/streaming/_page.mdx

Initiates a streaming conversation using cURL. This example shows the HTTP request structure for starting a stream.

```bash
curl --location "https://api.mistral.ai/v1/conversations" \
    --header 'Content-Type: application/json' \
    --header 'Accept: text/event-stream' \
    --header "Authorization: Bearer $MISTRAL_API_KEY" \
    --data '{
    "inputs": "Who is Albert Einstein?",
    "stream": true,
    "agent_id": "ag_06811008e6e07cb48000fd3f133e1771"
}'
```

--------------------------------

### Install Mistral AI Python SDK

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/developer/first-api-request.md

Install the Mistral AI Python SDK using pip. This is a prerequisite for running Python examples.

```bash
pip install mistralai
```

--------------------------------

### Example Fine-tuning Instance with Instructions

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/_guides/finetuning_sections/_02_prepare_dataset/page.mdx

Shows a single data instance formatted for fine-tuning, where the user message includes instructions for extracting structured information.

```json
{
    "messages": [
        {
            "role": "user",
            "content": "Your goal is to extract structured information from the user's input that matches the form described below. When extracting information please make sure it matches the type information exactly...Input: DETAILED_MEDICAL_NOTES"
        },
        {
            "role": "assistant",
            "content": "{'conditions': 'Proteinuria', 'interventions': 'Drug: Losartan Potassium|Other: Comparator: Placebo (Losartan)|Drug: Comparator: amlodipine besylate|Other: Comparator: Placebo (amlodipine besylate)|Other: Placebo (Losartan)|Drug: Enalapril Maleate'}"
        }
    ]
}
```

--------------------------------

### Response Example for Get User

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/admin/users/page.mdx

This is an example of a successful JSON response when retrieving user details. The 'uuid' field uniquely identifies the user.

```json
{
  "email": null,
  "first_name": null,
  "last_name": null,
  "uuid": "019b2bd7-96e7-7219-8c0b-45a73da50088"
}
```

--------------------------------

### Response Example for Get Spend Limits

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/admin/billing/page.mdx

This is an example of a successful response when retrieving spend limits. It includes information about monthly limits and currency.

```json
{
  "limits": {
    "completion": {
      "monthly_limit_reached": false
    },
    "currency": "ipsum eiusmod",
    "last_payment_failure": false,
    "last_payment_failure_protection": null
  }
}
```

--------------------------------

### Configuration Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/studio-api/connectors/confirmation/page.mdx

Example of how to configure tools to require confirmation.

```APIDOC
## Configuration

Add `requires_confirmation` to the `tool_configuration` of any Connector or built-in tool, and list the tool names that require approval:

```json
[
  {
    "type": "connector",
    "connector_id": "gmail",
    "tool_configuration": {
      "requires_confirmation": ["gmail_search"]
    }
  },
  {
    "type": "web_search_premium",
    "tool_configuration": {
      "requires_confirmation": ["web_search", "news_search"]
    }
  }
]
```
```

--------------------------------

### Run Python Quickstart Script

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/developer/first-api-request.md

Execute the Python script that sends a request to the Mistral API. This assumes the script is saved as quickstart.py.

```bash
python quickstart.py
```

--------------------------------

### Response Example for Get Text Content

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/libraries/documents/page.mdx

This is an example of a successful JSON response when retrieving text content from a document. It contains the extracted text.

```json
{
  "text": "Example text."
}
```

--------------------------------

### Launch Vibe CLI

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/getting-started/quickstarts/vibe-code/install-cli/page.mdx

Open a terminal in your project directory and run the 'vibe' command to launch the CLI. This will initiate a setup wizard for API key registration.

```bash
vibe
```

--------------------------------

### Get Prompt Version using cURL

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/prompts/page.mdx

An example using cURL to make a GET request to the Get Prompt Version endpoint. Replace placeholders with your actual prompt ID, version, and API key.

```curl
curl https://api.mistral.ai/v2/prompts/{prompt_id}/versions/{version} \
 -X GET \
 -H 'Authorization: Bearer YOUR_APIKEY_HERE'

```

--------------------------------

### Example Transcription Response with Diarization

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/studio-api/audio/speech_to_text/offline_transcription/passing_transcription_audio_url_tab/diarization_tab/_page.mdx

This is an example of a transcription response that includes speaker diarization. Each segment is tagged with a speaker ID, start, and end times.

```json
{
    "segments": [
        {"text": "It falls on each of us to be guardians of our democracy, to embrace the joyous task we've been given to continually try to improve this great nation of ours.", "start": 165.5, "end": 174.8, "speaker_id": "speaker_1", "type": "transcription_segment"},
        {"text": "Because for all our outward differences, we all share the same proud title, citizen.", "start": 175.7, "end": 181.9, "speaker_id": "speaker_1", "type": "transcription_segment"},
        {"text": "It has been the honor of my life to serve you as President.", "start": 183.3, "end": 186.1, "speaker_id": "speaker_1", "type": "transcription_segment"},
        {"text": "Eight years later, I am even more optimistic about our country's promise, and I look forward to working along your side as a citizen for all my days that remain.", "start": 187.1, "end": 197.5, "speaker_id": "speaker_1", "type": "transcription_segment"},
        {"text": "Thanks, everybody.", "start": 198.7, "end": 199.8, "speaker_id": "speaker_1", "type": "transcription_segment"},
        {"text": "God bless you, and God bless the United States of America.", "start": 200.1, "end": 203.5, "speaker_id": "speaker_1", "type": "transcription_segment"}
    ],
    "usage": {
        "prompt_audio_seconds": 203,
        "prompt_tokens": 7,
        "total_tokens": 938,
        "completion_tokens": 931,
        "request_count": 1,
        "num_cached_tokens": 0
    },
    "type": "transcription.done"
}
```

--------------------------------

### Start Local Development Server

Source: https://github.com/mistralai/platform-docs-public/blob/main/README.md

Launch a local development server for live previewing changes. The server typically runs at http://localhost:3000.

```bash
pnpm dev
```

--------------------------------

### List Deployments Output Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/workflows/managing-workflows-in-production/deployments.md

Example JSON output for listing deployments, showing deployment ID, name, active status, and creation/update timestamps.

```json
{
  "deployments": [
    {
      "id": "019d2585-21fc-7063-90ae-31a283c784a1",
      "name": "invoice-service",
      "is_active": true,
      "created_at": "2026-03-25 15:02:55.234755+00:00",
      "updated_at": "2026-04-02 17:42:34.960883+00:00"
    }
  ]
}
```

--------------------------------

### Response Example for Get Index Summaries

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/rag/search_indexes/page.mdx

This JSON structure represents a successful response from the GET /v1/rag/indexes/summary endpoint, detailing information about each search index.

```json
[
  {
    "created_at": "2025-12-17T10:25:07.818693Z",
    "creator_id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
    "document_count": 87,
    "id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
    "index": {
      "k8s_cluster": "production-cluster",
      "k8s_namespace": "search",
      "schemas": [
        {
          "document_count": null,
          "id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
          "name": "My resource"
        }
      ],
      "vespa_instance_name": "mistral-search"
    },
    "modified_at": "2025-12-17T10:41:03.469341Z",
    "name": "My resource",
    "status": "online"
  }
]
```

--------------------------------

### First Prompt Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/vibe/code/vs-code-extension/install-authenticate/page.mdx

Try this simple prompt to verify that the Mistral Vibe extension is set up correctly and can access your file context.

```text
Explain the currently open file.
```

--------------------------------

### Example Response for Get Invite

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/admin/users/page.mdx

This is an example of a successful response (200 OK) when retrieving pending organization invitations. It shows the structure of the invitation data returned.

```json
[
  {
    "email": null,
    "invite_uuid": "019b2bd7-96e7-7219-8c0b-45a73da50088",
    "role": "A",
    "roles": [
      "A"
    ]
  }
]
```

--------------------------------

### Example Response for Get Index Schema File

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/rag/search_indexes/page.mdx

This example shows a successful JSON response when retrieving a schema file. The 'content' field can be a string or null.

```json
Example content.
```

--------------------------------

### Complete Deal Analysis Workflow Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/workflows/building-workflows/durable_agents.md

A comprehensive example demonstrating a workflow with custom activities, agent coordination, and handoffs using RemoteSession. Includes agent definitions and workflow execution.

```python
import asyncio
import mistralai.workflows as workflows
import mistralai.workflows.plugins.mistralai as workflows_mistralai
from mistralai.client.models import TextChunk

@workflows.activity()
async def calculate_risk_score(deal_type: str, amount: float) -> dict:
    """Calculate financial risk score for a deal.

    Args:
        deal_type: The type of deal being analyzed
        amount: The monetary amount of the deal
    """
    risk_score = min(100.0, amount / 10000.0)
    risk_factors = []
    if amount > 100000:
        risk_factors.append("High value transaction")
    return {"risk_score": risk_score, "risk_factors": risk_factors}

@workflows.workflow.define(name="deal_analysis_workflow")
class DealAnalysisWorkflow:
    @workflows.workflow.entrypoint
    async def entrypoint(self, deal_request: str) -> dict:
        """Analyze a deal request.

        Args:
            deal_request: The deal request to analyze
        """
        session = workflows_mistralai.RemoteSession()

        # Risk assessment agent
        risk_agent = workflows_mistralai.Agent(
            model="mistral-medium-latest",
            name="risk-agent",
            description="Analyzes financial risk of deals",
            instructions="Use the risk calculation tool to assess deal risk.",
            tools=[calculate_risk_score],
        )

        # Main coordinator agent
        coordinator = workflows_mistralai.Agent(
            model="mistral-medium-latest",
            name="deal-coordinator",
            description="Coordinates deal analysis",
            instructions="Analyze the deal request and hand off to specialists.",
            handoffs=[risk_agent],
        )

        outputs = await workflows_mistralai.Runner.run(
            agent=coordinator,
            inputs=deal_request,
            session=session,
        )

        analysis = "\n".join([
            output.text for output in outputs
            if isinstance(output, TextChunk)
        ])

        return {"analysis": analysis}

if __name__ == "__main__":
    asyncio.run(workflows.run_worker([DealAnalysisWorkflow]))
```

--------------------------------

### Create a Web Search Agent in Python

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/llms-full.txt

This example demonstrates creating an agent with the built-in web search tool. It configures the agent to use `web_search` for up-to-date information and sets a specific temperature and top_p for completions.

```python
websearch_agent = client.beta.agents.create(
    model="mistral-medium-2505",
    description="Agent able to search information over the web, such as news, weather, sport results...",
    name="Websearch Agent",
    instructions="You have the ability to perform web searches with `web_search` to find up-to-date information.",
    tools=[{"type": "web_search"}],
    completion_args={
        "temperature": 0.3,
        "top_p": 0.95,
    }
)
```

--------------------------------

### Request Body Example for Starting a Conversation

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/conversations/page.mdx

This JSON object shows the structure for the request body when starting a new conversation. It includes the 'inputs' field for your initial prompt.

```json
{
  "inputs": "Example input."
}

```

--------------------------------

### Manual Install Vibe CLI with pip

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/getting-started/quickstarts/vibe-code/install-cli.md

Alternative manual installation method using `pip`.

```bash
pip install mistral-vibe
```

--------------------------------

### Get Workflow Registrations Response Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/workflows/page.mdx

This is an example of a successful JSON response when retrieving workflow registrations. It includes details about next cursor, workflow registrations, and workflow versions.

```json
{
  "next_cursor": null,
  "workflow_registrations": [
    {
      "definition": {
        "input_schema": [
          null
        ]
      },
      "id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
      "workflow_id": "019b2bd7-96e7-7219-8c0b-45a73da50088"
    }
  ],
  "workflow_versions": [
    {
      "definition": {
        "input_schema": [
          null
        ]
      },
      "id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
      "workflow_id": "019b2bd7-96e7-7219-8c0b-45a73da50088"
    }
  ]
}
```

--------------------------------

### Check Node.js Version

Source: https://github.com/mistralai/platform-docs-public/blob/main/scannability_test_report.md

Command to check the installed Node.js version. Required for TypeScript examples.

```bash
node --version
```

--------------------------------

### Create Agent with Web Search Tool

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/studio-api/agents/handoffs/page.mdx

Example of creating an agent that can search online for information. This agent is configured with a 'web_search' tool.

```bash
curl --location "https://api.mistral.ai/v1/agents" \
     --header 'Content-Type: application/json' \
     --header 'Accept: application/json' \
     --header "Authorization: Bearer $MISTRAL_API_KEY" \
     --data '{ \
     "model": "mistral-large-latest", \
     "name": "web-search-agent", \
     "description": "Agent that can search online for any information if needed", \
     "tools": [{"type": "web_search"}] \
  }'
```

--------------------------------

### Response Example (JSON)

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/libraries/page.mdx

This is an example of a successful response when retrieving library details. It includes information such as creation date, ID, name, and size.

```json
{
  "chunk_size": null,
  "created_at": "2025-12-17T10:25:07.818693Z",
  "id": "019b2bd7-96e7-7219-8c0b-45a73da50088",
  "name": "My resource",
  "nb_documents": "426",
  "owner_id": null,
  "owner_type": "User",
  "total_size": "6436546",
  "updated_at": "2025-12-17T10:41:03.469341Z"
}

```

--------------------------------

### Start a Conversation in Python

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/api/endpoint/beta/conversations/page.mdx

This Python snippet demonstrates how to start a conversation and get the conversation ID using the Mistral AI Python SDK. Make sure the MISTRAL_API_KEY environment variable is configured.

```python
import os
from mistralai import Mistral

client = Mistral(api_key=os.environ["MISTRAL_API_KEY"])

conversation = client.beta.conversations.start(
    inputs="Help me summarize this document.",
    stream=False,
)

print(conversation.id)

```

--------------------------------

### Initialize Mistral Client (V2 SDK)

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/knowledge-rag/rag_quickstart.md

Installs and imports necessary packages for RAG, then initializes the Mistral client with your API key. Use this for setting up your environment.

```python
from mistralai.client import Mistral
import requests
import numpy as np
import faiss
import os
from getpass import getpass

api_key = getpass("Type your API key")
client = Mistral(api_key=api_key)
```

--------------------------------

### Workflow Failure Example

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/vibe/work/workflows.md

This example shows the typical output when a Workflow fails due to access restrictions, such as an account tier limitation. The output includes a 'started' message, the error reason, and a 'failed' message.

```text
Workflow started: vibe-nuage
Your account tier (free) does not have access to this feature
Workflow failed: vibe-nuage
```

--------------------------------

### Direct Example in Prompt

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/conversations/chat-completion/prompting.md

A concise example showing the input and expected output format within a prompt. This illustrates a direct interaction pattern.

```plaintext
[...]
# Examples
Input: Hello, how are you?
Output: {"language_iso": "en"}
[...]
```

--------------------------------

### Start Conversation with Moderation

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/conversations/moderation.md

This example demonstrates how to start a conversation with moderation enabled by specifying guardrails directly in the request. It shows how to set custom thresholds for specific categories and define the action to take if a guardrail is violated.

```APIDOC
## POST /v1/conversations

### Description
Starts a new conversation with specified inputs, model, and guardrails for content moderation.

### Method
POST

### Endpoint
/v1/conversations

### Parameters
#### Request Body
- **model** (string) - Required - The model to use for the conversation.
- **inputs** (array) - Required - An array of message objects representing the conversation history.
- **guardrails** (array) - Optional - An array of guardrail configurations.
  - **blockOnError** (boolean) - Optional - If true, the request will be blocked if a guardrail error occurs.
  - **moderationLlmV2** (object) - Optional - Configuration for the v2 moderation LLM.
    - **customCategoryThresholds** (object) - Optional - A map of category names to their threshold values.
      - **sexual** (number) - Optional - Threshold for sexual content.
      - **selfharm** (number) - Optional - Threshold for self-harm content.
    - **ignoreOtherCategories** (boolean) - Optional - If true, other categories will be ignored.
    - **action** (string) - Optional - The action to take when a guardrail is violated (e.g., "block").

### Request Example
```json
{
  "model": "mistral-small-latest",
  "inputs": [{ "role": "user", "content": "How far is the moon from Earth?" }],
  "guardrails": [
    {
      "block_on_error": true,
      "moderation_llm_v2": {
        "custom_category_thresholds": {
          "sexual": 0.1,
          "selfharm": 0.1
        },
        "ignore_other_categories": false,
        "action": "block"
      }
    }
  ]
}
```

### Response
#### Success Response (200)
- **object** (string) - Type of the response object.
- **conversation_id** (string) - The ID of the conversation.
- **outputs** (array) - An array of message objects representing the model's response.
- **usage** (object) - Token usage information.
- **guardrails** (array) - Information about the applied guardrails and their results.
```

--------------------------------

### Vibe Code Slack Task Examples

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/content/en/docs/vibe/code/vibe-code-web/slack-integration/page.mdx

Examples of how to prompt Vibe Code in Slack for different development tasks. These prompts guide Vibe to perform specific actions like fixing bugs or updating documentation.

```text
@Vibe use the customer report above to find the likely bug. Open a PR if you can reproduce or identify a clear fix.
```

```text
@Vibe investigate this alert and check whether there is a code or config fix. Start with a summary, then open a PR only if the change is low risk.
```

```text
@Vibe implement the copy/config change we agreed on above and open a PR.
```

```text
@Vibe add a regression test for the edge case described in this thread. Keep the implementation unchanged unless the test exposes a bug.
```

```text
@Vibe update the docs based on the behavior described above. Open a PR with the doc-only change.
```

```text
@Vibe refactor the helper discussed above. Keep behavior unchanged and run the relevant tests before opening a PR.
```

--------------------------------

### Fetch and Build API Documentation

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/scripts/api/README.md

Use this sequence of commands to pull the latest API specification, build the documentation (including MDX generation and postprocessing), and audit for changes. This is typically followed by a review step.

```bash
pnpm api:fetch                  # pull latest spec
pnpm api:build                  # apply + docs-md + postprocess (regenerates MDX)
pnpm api:audit                  # parse MDX, find lorem hits, write draft
```

--------------------------------

### Document Annotation Prompt Examples

Source: https://github.com/mistralai/platform-docs-public/blob/main/public/studio-api/document-processing/annotations.md

Provides examples of system prompts for document annotation in Python, TypeScript, and cURL. These prompts guide the model on extracting specific information like language, chapter titles, and URLs from documents.

```python
document_annotation_prompt = """
Extract the following from the provided PDF document:
- Language (e.g., "English")
- All chapter/section titles (e.g., ["Abstract", "1 Introduction"])
- All URLs (e.g., ["https://example.com"])
Be precise and include only exact matches.
"""
```

```typescript
const documentAnnotationPrompt = `
Extract the following from the provided PDF document:
- Language (e.g., "English")
- All chapter/section titles (e.g., ["Abstract", "1 Introduction"])
- All URLs (e.g., ["https://example.com"])
Be precise and include only exact matches.
`;
```

```bash
"document_annotation_prompt": "Extract the following from the provided PDF document:\n- Language (e.g., \"English\")\n- All chapter/section titles (e.g., [\"Abstract\", \"1 Introduction\"])\n- All URLs (e.g., [\"https://example.com\"])\nBe precise and include only exact matches."
```

--------------------------------

### Promote Draft Proposals

Source: https://github.com/mistralai/platform-docs-public/blob/main/src/scripts/api/README.md

Moves draft API examples into the versioned `api-examples.yaml` file, preparing them for promotion.

```bash
pnpm api:promote
```