Home/HighLevel/Features/Workflows/HighLevel workflow guide for using Mistral AI actions

Workflows and Automation

HighLevel workflow guide for using Mistral AI actions

In short

Mistral AI workflow actions let HighLevel call language, embedding, and vision models directly from automations. After configuring a prompt, text, or image, the model returns a result that becomes a workflow variable for later steps such as messaging, field updates, or conditional logic.

AreaWorkflows · Workflow Actions
Used forEcommerce store, Legal practice, Real estate agency
Setup steps7 steps in 1 task
Last updatedNovember 17, 2026

01Who this is for

  • Founders who want to embed AI generated replies into their client communication without building separate services.
  • Marketing teams that need multilingual content creation on the fly for email and SMS campaigns.
  • Operations managers looking to extract data from uploaded receipts or IDs automatically.

02Mistral AI actions, explained

Mistral AI actions are native workflow steps that send prompts, text, or image URLs to Mistral models and receive the result as a workflow output variable. The integration works after a standard HighLevel trigger starts the automation; the AI step can be followed by any other action that accepts variables.

The actions include Create Chat Completion for text generation, Create Embeddings for vector conversion, and Analyze Image for vision tasks. Each action provides fields for prompts, system prompts, model selection, and optional tuning parameters. The output variable can be inserted into messages, stored in contact fields, or evaluated in conditional logic.

03How businesses use mistral AI actions

  • Ecommerce store. When a customer submits a review, the store adds a Create Chat Completion step that rewrites the review into a friendly reply. The generated text is then placed in a Send Email action to thank the buyer.
  • Legal practice. A client uploads a scanned contract image via a form. The workflow uses Analyze Image with a Pixtral model to extract key dates and parties, then updates custom fields on the client record automatically.
  • Real estate agency. Incoming inquiries are turned into embeddings with Create Embeddings, enabling the workflow to match the question to a pre indexed FAQ vector and send the best answer using Send SMS.

04The numbers and rules that matter

  • Embedding output dimension is 1024 dimensions.
  • Models include Large, Medium, Small, Ministral 8B/3B, Pixtral Large, Pixtral 12B.

05Before you start

  • A Mistral API key must be generated in console.mistral.ai.
  • The Mistral integration must be connected in HighLevel via Settings > Integrations > Mistral AI or from a workflow step.
  • The workflow must contain a trigger such as Customer Replied, Form Submitted, or Inbound Webhook.

06How to set up and use mistral AI actions in HighLevel

Set up Mistral AI in a HighLevel workflow

  1. Open Automation > Workflows and create a new workflow or edit an existing one.
  2. Add a Mistral AI step and click Connect Now, or go to Settings > Integrations > Mistral AI > Connect to authorize the integration.
  3. Select the required Mistral AI action, Create Chat Completion, Create Embeddings, or Analyze Image.
  4. Fill in the action fields such as Action Name, System Prompt, Prompt, Model, and any tuning parameters, inserting workflow variables where needed.
  5. Add a following step like Send SMS, Send Email, Update Contact Field, Custom Webhook, or If/Else and reference the Mistral output variable.
  6. Trigger a test event, run the workflow on a test record, and review the Run Logs for inputs and outputs.
  7. Click Publish to activate the workflow.

07What it looks like in HighLevel

08Settings that matter

  • Action Name. A label you assign to the step to identify its purpose in the workflow.
  • System Prompt. Defines the AI’s role, tone, or formatting rules for the model response.
  • Prompt. The specific task description that the model will execute; you can embed workflow variables.
  • Model. Chooses which Mistral model family (Large, Medium, Small, Ministral 8B/3B, etc.) the action will call.
  • Tuning Parameters. Controls output style, length, and other generation settings provided by the model.

09Things to know before you rely on it

  • Mistral AI actions are in Beta; models, fields, and behavior may change over time.
  • Each action consumes premium workflow action credits in addition to Mistral usage charges.
  • Mistral rate limits and usage quotas are enforced by your Mistral workspace, not by HighLevel.
  • Free Mistral plans may have stricter usage limits than paid plans.

10How it connects to the rest of HighLevel

  • Review HighLevel premium feature pricing for workflow credits Pricing
  • Start a free 30 day trial to test Mistral AI actions 30 day trial
  • Founders can access all workflow features through the Founder Plan Founder Plan

Michael's take

Michael Reimer
Michael ReimerFounder of CRM Pros, HighLevel Commander and Certified Admin

Mistral AI actions let you automate multilingual communication and extract information from images without writing code.

If your team only needs occasional text tweaks or you already have an integration that handles document parsing, you might prefer native HighLevel tools until the beta matures.

Get More InformationOpens on HighLevelPublished November 2, 2026 · Updated November 17, 2026

Common questions

Common questions about mistral AI actions.

What does Mistral AI do in a HighLevel workflow?

It sends the configured prompt, text, or image to a Mistral model and returns the response as a workflow output variable that can be used in later steps.

Do I need a separate Mistral account?

Yes. You must provide an API key from console.mistral.ai, and Mistral bills usage directly to that account.

Can I include workflow variables inside the prompts?

Both the System Prompt and the Prompt fields accept inline variables such as contact fields, custom fields, trigger payload values, and earlier step outputs.

Are the Mistral AI actions considered premium actions?

Yes. They consume premium workflow action credits at the standard automation rate, in addition to any Mistral usage fees.

What dimension does the embedding output have?

The vector produced by Create Embeddings contains 1024 dimensions and can be sent to an external vector database via a webhook.

Next step

Pick the features that matter. Then get the plan that unlocks them.

Explore the features you want, start with the five at the top, and upgrade through my link when you are ready. The bonuses and the team come with it.

The features directory explains every HighLevel feature in plain English. Browse it to see what the platform can do.

Back to the FeaturesSee the full details
⌃

Site preview

Explore the next page

This destination is reserved in the page inventory and is not built yet.