BrazeAI Agent Console

What it does

BrazeAI Agent Console is a dashboard inside the Braze platform where marketers define an agent's task, choose a model, and write instructions in a prompt. The agent takes in context you pass to it at runtime, such as user profile fields, catalog rows, Canvas context variables, and brand guidelines, and produces output such as generated message copy, sentiment scores, customer attributes, or catalog matches. Its outputs are stored in output variables that can drive personalized messages or route users through Canvas paths.

Teams use it to generate and localize content, score leads, translate messages, convert unstructured inputs like survey responses into structured fields, and adapt communications to customer behavior. Its distinct role is reasoning over one user's context at send time, not learning from results. Agents are invoked in Canvases or Catalogs, and the console offers testing, run logs, version history, and daily invocation limits.

Agent Console suits marketers already using Braze who need LLM-driven personalization inside Canvases and catalogs. Access requires Message or Action Credits, and agent runs consume credits. Agents only see data you explicitly pass in, return null if a run exceeds 20 seconds, and the Braze-provided default model uses Google Gemini.

What it can do

• Create custom LLM-powered agents with a task, model choice, and prompt instructions, without coding
• Deploy agents as Canvas steps to personalize messages and route users through paths
• Deploy agents in catalog fields to enrich product and profile data
• Convert unstructured inputs such as survey responses or chat messages into structured outputs
• Test agents with sample inputs and review per-run logs and version history
• Set daily invocation limits and fallback values for Canvas Step Agents

Works with

• OpenAI
• Anthropic
• Google Gemini
• Databricks Mosaic
• AWS
• Microsoft Foundry

Before you choose

• Agents do not learn from outcomes; use BrazeAI Decisioning Studio for learning-based optimization.
• Agents only receive data you explicitly configure; they do not search profiles or warn when data is missing.
• Default runs time out after 20 seconds, input is limited to 25 KB, and timeouts consume credits.

Listed under

• Coding
• MCP and agent tools

Get started

Action Credits are required; the docs direct users without credits to their account manager.

1. Confirm your account has Message or Action Credits, or contact your account manager for next steps.
2. Open Agent Console, define the agent's task, choose a model, and write instructions in the prompt.
3. Test the agent with sample inputs, review logs, and configure fallback values before deploying it in a Canvas or catalog.

Agent runs consume Action Credits, and the Platform Edition and credit allotment are set in the contract. Braze does not offer a free plan at this time. Confirm the credit cost per invocation with Braze before deploying at scale.

We couldn’t find published prices for BrazeAI Agent Console. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Braze Agents in Agent Console ↗www.braze.com
• Scale Smarter Engagement | BrazeAI™ Agents ↗www.braze.com
• Explore pricing and packaging options at Braze ↗www.braze.com
• Braze Customer Engagement Platform ↗www.braze.com
• AI Marketing & Personalization Tools | BrazeAI™ ↗www.braze.com
• Documentation ↗www.braze.com

Company website

braze.com: product, pricing and policy pages

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.