You can now easily identify which Marketplace App sent a message in Conversations, directly from the message details.

👀

We’ve made updates to help you better understand and troubleshoot failed email messages—without interrupting your workflow.

Why this matters

These improvements are designed to make debugging faster and your workflow smoother.

We’ve revamped the snapshot import experience to make it faster, clearer, and more intuitive—giving you greater control right from the start.

Try importing a snapshot using a share link to explore the new flow and custom naming experience.

Chat Widget configurations can now be included in Snapshots, allowing agencies to replicate widget setups across sub-accounts with proper handling of account-specific dependencies like WhatsApp numbers, Voice AI agents, and integrations.

Creating a Custom Product Details Page (PDP) is now faster and more intuitive.

Earlier, enabling the “Create Custom Product Details Page” option would generate a blank page—forcing users to rebuild the layout from scratch.

With this update, new Custom PDPs automatically inherit the design of the existing Default Product Details Page, allowing users to start from a fully structured, production-ready layout.

Images:

Building stable, future-proof integrations just got easier. We’ve introduced API Versioning for our public documentation, enabling developers to reliably access and work with specific API versions without worrying about unexpected breaking changes.

Open API documentation → Select the desired version from the version switcher → Reference endpoints and schemas for that version → Build with confidence

We are working on the next milestone — v3 of our APIs, which will introduce enhanced capabilities, improved performance, and new endpoints for advanced use cases.

Select version by switching via toggle at top of the API documentation:

Nudge highlighting user that they are in older version:

We’ve expanded app update versioning to include AI Agent modules, including Conversation AI and Voice AI templates.

Earlier, app versioning supported the release flow for app updates, but AI Agent module changes were not fully tied to app versions. With this update, AI Agent template changes are now version-controlled, making releases safer, clearer, and easier to manage.

This update gives developers better control over how AI Agent changes are released and gives users more visibility into what changed before updating an installed app.

AI Agent updates are now easier to track, review, and adopt through Marketplace release notes and version updates.

In case of any further questions, please reach out to marketplace@gohighlevel.com

AI is most useful when it sits inside the rest of your automation, not as a separate manual step. The Mistral AI integration lets your workflows call Mistral language models, generate embeddings for semantic search, and analyze images — using your own Mistral API key, with the response available as a workflow variable for any later step.

This means workflows can reply to a customer in their own language, look up the closest match in a knowledge base, or pull structured data out of an uploaded receipt — without sending the contact off to a third-party tool.

Mistral AI is a French AI lab that builds open-weight and frontier language models. The integration ships three actions:

All three actions are premium actions

Before any Mistral AI action can run, the integration has to be connected. The connection allows your workflows to securely call Mistral using your API key.

Connect via the Workflow Builder (recommended)

Connect via Settings (alternative path)

There are no Mistral AI triggers in this iteration. Mistral AI is currently action-only — workflows can call Mistral models from any step, but cannot be initiated by a Mistral event. Trigger support may be added in a later release.

This walkthrough creates an automated multilingual reply. The workflow listens for an inbound message, asks Mistral to draft a friendly response in the contact’s language, and sends it back over email.

Open or create a workflow. Click Add to insert a step.

This example pulls structured data out of an uploaded receipt and writes it to the contact record.

Inbound message → Mistral AI Create Chat Completion → Send Email/SMS

2. Semantic FAQ matching for inbound questions

Inbound email/customer replied → Create Embeddings → Vector lookup → Send templated reply

3. Receipt and document extraction from images

Inbound webhook -> Image upload (form / inbound MMS) → Analyze Image (Vision) → Update contact Create opportunity

Q: What is Mistral AI?

Mistral AI is an AI lab that develops open-weight and frontier language and vision models. The integration gives Workflows direct, governed access to those models using your own API key.

Q: Which Mistral models are available in the model dropdown?

Models from the Mistral Large, Medium, Small, Ministral and Magistral families. The list is refreshed periodically as Mistral retires legacy versions and ships new ones.

Q: Do I need a separate Mistral account?

Yes. Bring your own API key from console.mistral.ai. Inference is billed by Mistral directly based on your account tier and chosen model.

Q: Are Mistral AI actions premium workflow actions?

Yes. All three actions consume premium action credits at the standard automation rate, in addition to Mistral usage charges.

Q: Can I use workflow variables inside prompts?

Yes. Both the System Prompt and the Prompt accept inline variables — contact fields, custom fields, trigger payload values, and the outputs of earlier workflow steps.

Q: Where is the AI-generated response stored?

It is exposed as a workflow output variable on the action. Use it in messages, write it to custom fields, or pass it to downstream steps. It is not persisted as a separate record by default.

Q: What happens if Mistral returns an error or times out?

The action surfaces the error to the workflow execution log, and the workflow continues based on its configured branching. Always test the action before publishing, and add a fallback branch if reliability matters.

Q: Can the generated response be used in SMS or email messages?

Yes. The output variable can be inserted into Send SMS, Send Email, internal notifications, and any other messaging action.

Q: What is the embedding output dimension?

1024 dimensions. The vector is suitable for direct use with major vector databases. Pass it to a webhook action that writes to your vector store.

Q: What image formats does Analyze Image accept?

Standard PNG and JPEG via direct URL, including images stored in Media Storage that are referenced by file URL. Images are processed at native resolution by the underlying vision model.

Q: Are these actions production-ready?

All three actions are released as Beta. They are safe to use in live workflows; expect incremental UX changes — for example, additional models added to the dropdown — as the integration moves out of Beta.

We’ve added the ability for agencies to enable temporary support access to their accounts from within the platform.

What’s Included

New Support Access section under:

Agency View → Settings → Company → Advanced Settings

Users can:

Enable support access

Define a time-bound duration

Extend or revoke access anytime

Why This Was Built

This is a security and compliance-driven update to ensure:

Help article :