Skip to content
Novu logoNovu

Changelog

Product updates, improvements, and fixes

Follow us on X

All changelog posts

  • Manage agents and conversations through the Novu MCP server

    Create and update agents, connect channels, and inspect conversation activity from any MCP-compatible AI tool.

    Authors:
    Nikita GrossmanVictor Yakubu
    Nikita G., Victor Y.
    Novu MCP Server

    Agent and conversation tools are now available on the Novu MCP server. You can manage the conversational side of Novu from the same AI tool you use to create workflows, inspect subscribers, and troubleshoot notification delivery.

    Ask your assistant to create or update an agent, connect a channel, list recent conversations, or inspect the activity timeline for a specific thread. Together with the existing workflow tools, you can also assign agents to existing workflows without switching between your editor and the Novu Dashboard.

    Create and configure agents from a prompt

    The new agent tools cover the core setup and management flow:

    ToolWhat it does
    create_agentCreates a Novu agent.
    get_agentsLists agents with cursor-based pagination.
    get_agentRetrieves one agent by its public identifier.
    update_agentUpdates an agent's name, status, bridge URL, or behavior.
    connect_agentConnects a channel to an existing agent.

    For example, you can ask:

    Create a managed Novu agent called Support Bot that answers product questions briefly, then connect Slack to it.

    Your AI tool discovers the required Novu operations and calls them in sequence. Agent tools use the agent's public identifier, such as support-bot, rather than its internal database ID.

    Inspect conversations where you build

    Two conversation tools bring agent debugging into the same workflow:

    • get_conversations lists agent conversations and supports optional filters.
    • get_conversation_activities returns the activity timeline for a conversation.

    You can move from an agent to its latest conversation and inspect the full thread activity with one prompt:

    Show recent conversations for the support-bot agent, then inspect the activity timeline for the latest one.

    This makes it easier to investigate a failed turn, review what happened around a tool approval, or trace a conversation back to its channel without starting in the Dashboard.

    Use the tools in the right environment

    OAuth connections default to the Development environment. Ask the assistant to list your environments, then pass the target environment when you need to work in Production or another environment. API keys remain scoped to the environment they belong to.

    The account-management MCP server is available for Novu Cloud at:

    • US: https://mcp.novu.co/
    • EU: https://eu.mcp.novu.co/

    OAuth is the recommended authentication method for interactive clients. A Novu Cloud API key is available as a fallback for clients that do not support OAuth.

    The hosted endpoints do not connect to self-hosted Novu installations. The conversation tools also require Novu Cloud with Conversations enabled.

    Connect the Novu MCP server to Cursor, Claude Code, ChatGPT, VS Code, Codex, or another MCP-compatible tool to get started.

  • Agent-Assigned Workflows

    Turn a workflow notification into the first turn of an agent conversation, with the original message and trigger payload already in context.

    Authors:
    Nikita GrossmanVictor Yakubu
    Nikita G., Victor Y.
    Cover image for Agent-Assigned Workflows

    You can now assign an agent to a workflow from the Novu Dashboard. Workflow messages are sent through the agent's connected channels, and when a subscriber replies, Novu routes the reply to that agent with the originating workflow, message, and trigger payload already in context.

    This closes the gap between notifications and conversations. A workflow can send “Your order has shipped,” the subscriber can reply “Can I change the delivery address?”, and the agent receives the order data from the original trigger without a separate correlation layer in your application.

    Turn a notification into a conversation

    Enable Send & reply via agent in the workflow editor and select the agent that should handle replies. Novu associates the outgoing workflow message with that agent and delivers it through the agent's applicable channel connection.

    When the subscriber replies, Novu maps the response back to the sent notification. Depending on the channel, this uses the email reply token, the Slack thread, a quoted message on WhatsApp, Telegram, or Microsoft Teams, or the existing one-to-one conversation history on iMessage.

    The original workflow context is attached when the conversation starts. Later replies reuse that context, so your application does not need to reload and attach the same notification data on every turn.

    For example, an order-support workflow could provide the agent with the rendered shipping message and a payload containing the order ID, tracking number, and delivery address. The agent can then use those values when it handles a follow-up question.

    Send through the agent's connected channels

    An assigned workflow can send through agent-linked Slack, Microsoft Teams DM, WhatsApp, Telegram, iMessage through Sendblue, and email integrations when the subscriber is reachable on that connection.

    Email can use an agent-specific sender and optional reply-to address, so replies return to the agent instead of a no-reply inbox.

    If the selected agent does not have an applicable channel connection, Novu falls back to the provider integration configured in the environment and adds a warning to the step. The assignment does not silently drop the delivery. A Chat step is skipped only when the subscriber has no reachable Chat channel, which matches the existing delivery behavior.

    Give custom code and managed agents the workflow context

    Custom code agents receive a typed ctx.notification field when a conversation starts from an assigned workflow:

    import { isFromWorkflow } from '@novu/framework';
    import { orderShipped } from './workflows/order-shipped';
    
    if (isFromWorkflow(ctx.notification, orderShipped)) {
      const { orderId, trackingNumber } = ctx.notification.payload;
    }

    When the workflow defines a payload schema, ctx.notification.payload uses that schema for type inference. The field is null when the conversation started with an inbound message instead of a workflow notification.

    Managed agents receive the same originating workflow and payload data automatically as conversation context. Novu adds a short assistant message that identifies the workflow, followed by the payload as JSON. There is no handler to update.

    The Agent Conversations timeline in the Dashboard also shows the originating workflow notification, so you can trace how the conversation started alongside the replies that followed.

    Override the assigned agent for one trigger

    The workflow assignment is the default. You can route a single execution to another agent by passing its public identifier as agentId:

    await novu.trigger({
      workflowId: 'order-shipped',
      to: 'subscriber-123',
      payload: {
        orderId: 'order-456',
        trackingNumber: '1Z999AA10123456784',
      },
      agentId: 'order-support-agent',
    });

    Omit agentId to use the agent assigned in the workflow. Pass agentId: null when that execution should send without agent reply routing.

    Open a workflow in the Dashboard and enable Send & reply via agent to get started. See the AI SDK quickstart, Trigger event API, and Agent Conversations documentation for the related APIs and runtime behavior.

  • Provider content overrides in the Dashboard

    Save provider-native JSON on Chat and Tool workflow steps, then version and promote it with the rest of the workflow.

    Authors:
    George DjabarovVictor Yakubu
    George D., Victor Y.
    Provider content overrides in the dashboard

    Provider content overrides can now be configured directly in the Novu Dashboard. Use them when a workflow needs fields from a provider's native API that the shared step editor does not expose, such as Slack Block Kit, WhatsApp templates, or PagerDuty incident fields.

    Previously, these payloads had to be added to every trigger() call in application code. You can now save the JSON once on the workflow step. It applies to every trigger and moves from development to production with the workflow.

    Keep provider-native content with the workflow

    Open a Chat or Tool step and select a connected provider from the content source dropdown. Provider tabs only appear when that provider has an active integration in the current environment.

    The override is stored as a JSON object on the step. It supports Liquid variables, so the payload can still use workflow and subscriber data:

    {
      "blocks": [
        {
          "type": "section",
          "text": {
            "type": "mrkdwn",
            "text": "Deploy *{{payload.status}}* for {{subscriber.firstName}}"
          }
        },
        {
          "type": "actions",
          "elements": [
            {
              "type": "button",
              "text": { "type": "plain_text", "text": "View run" },
              "url": "{{payload.runUrl}}"
            }
          ]
        }
      ]
    }

    The Dashboard also shows a preview of the merged payload before you save the workflow.

    Get validation where a provider schema is available

    Schema-backed editors help you catch malformed provider payloads before delivery. Slack overrides include autocomplete and validation against supported chat.postMessage fields, including nested Block Kit elements. The editor also links to Slack's Block Kit Builder.

    WhatsApp Business overrides provide autocomplete and validation for message shapes such as templates, interactive messages, media, and text. The same override interface is available for Tool providers, including PagerDuty, Opsgenie, Grafana, and Tool Webhook.

    When a provider does not have a schema-backed editor, the Dashboard provides a free-form JSON field with a warning and a link to the provider's documentation. Novu passes that object to the provider, so test the workflow with a real trigger before relying on it.

    Combine default content and overrides predictably

    The shared step body remains the default content. Novu merges the saved override into the provider request. If the override omits the provider's primary content field, such as text for Slack or text.body for WhatsApp, Novu fills it from the rendered step body.

    Overrides follow this precedence order:

    1. Dashboard override saved on the step.
    2. Workflow-level provider override passed at trigger time.
    3. Step-level provider override passed at trigger time.

    The higher-priority value wins when the same field appears at multiple levels. Arrays replace the lower-priority array as a whole. They are not merged by index.

    Some Slack chat.postMessage fields do not apply when the integration delivers through an incoming webhook. Check the Slack setup guide before relying on fields such as thread or sender metadata.

    Use Rich Chat for shared content and overrides for provider-specific fields

    The Rich Chat editor and provider content overrides work together. Use the block editor when you want to create one structured message and let Novu render it for each connected chat provider. Use an override when one provider needs its own JSON format or a field outside the shared card model.

    Provider content overrides are available for Tool steps and are rolling out gradually for Chat steps. If the provider dropdown does not appear on a Chat step yet, trigger-time provider overrides continue to work.

    Read the Chat provider content overrides overview, configure Slack overrides in the Dashboard, or review the provider override precedence rules.

  • Rich chat editor for workflow Chat steps

    Build structured chat notifications with text, images, lists, link buttons, variables, and provider-aware previews from the workflow editor.

    Authors:
    Paweł TymczukVictor Yakubu
    Paweł T., Victor Y.
    Rich chat editor for workflow Chat steps

    Workflow Chat steps now include a visual block editor. Instead of limiting every workflow-sent chat notification to one text body, you can compose structured messages for approvals, reports, summaries, prompts, and calls to action from the Dashboard.

    Novu converts the shared message into a provider-specific format at delivery time. You author the content once, then inspect how it may appear on each chat provider connected to the environment.

    Rich chat editor

    Build structured Chat steps visually

    Type / or use the add control to insert blocks into a Chat step. The first release supports:

    • Text and images.
    • Dividers, blockquotes, hard breaks, and numbered or bulleted lists.
    • Link buttons with variable-backed labels and redirect URLs.
    • Repeat and Digest blocks for content produced by earlier workflow steps.

    Dashboard buttons open a URL. They do not route a reply to an agent or call back into your application.

    Preview each provider before publishing

    The preview panel uses the same shared card model as delivery. Switch between providers configured in the current environment to inspect an approximate rendering for each one.

    Novu maps the card into provider-specific formats for Slack, Microsoft Teams, Telegram, and WhatsApp. Providers without a matching rich layout receive a compatible Markdown or text representation. When a provider cannot represent a block or button fully, the preview shows a compatibility warning so you can simplify the message or add a provider content override.

    Provider previews are approximate. The delivered message can still differ because each platform controls its own spacing, layout, and supported features.

    Existing text workflows keep their behavior

    Existing Chat steps that use plain text or Liquid continue to use the Text editor and deliver as before. New empty Chat steps open in the Block editor by default.

    You can switch between the Block and Text editors, but changing modes replaces the content stored for the current mode. Copy any content you need before switching while editing an existing workflow.

    Liquid variables and translations continue to work. Use the Text editor for templates that depend heavily on Liquid conditions or loops. Use the Block editor when the message benefits from a structured layout that can be shared across providers.

    Return the same card model from Framework

    Code-first workflows can return a card from step.chat() instead of a text-only body:

    Return the same card model from Framework
    import { Actions, Card, CardText, Divider } from '@novu/framework';
    
    await step.chat('deploy-complete', async () => ({
      card: Card({
        title: 'Deploy complete',
        children: [
          CardText('All checks passed.'),
          Divider(),
          Actions([
            {
              type: 'link-button',
              id: 'view-deploy',
              label: 'View deploy',
              url: 'https://example.com/deploys/123',
            },
          ]),
        ],
      }),
    }));

    Existing body returns remain valid. When a Chat step returns both body and card, Novu uses the card.

    Add provider-native JSON when the shared card is not enough

    The block editor covers shared rich content. Provider content overrides remain available beside Default content when you need exact provider JSON, such as a full Slack Block Kit payload or a WhatsApp template.

    Read how to write Chat templates in the Dashboard or see the Framework Chat step for code-first workflows.

  • Tool steps for on-call services and webhooks

    Send workflow notifications to subscriber-specific PagerDuty, Opsgenie, Grafana IRM/OnCall, or custom HTTP endpoints through a first-class Tool channel.

    Authors:
    George DjabarovVictor Yakubu
    George D., Victor Y.
    Tool channel steps for on-call services and webhooks

    Tool is now a first-class channel in Novu workflows, alongside Email, SMS, Chat, and Push. Add a Tool step when a notification needs to create an on-call alert or reach an HTTP endpoint instead of a person-facing messaging channel.

    The first Tool providers are PagerDuty, Opsgenie, Grafana, and Tool Webhook.

    Tool channel step

    Route alerts to each subscriber's on-call service

    PagerDuty, Opsgenie, and Grafana use subscriber-specific channel endpoints. Each subscriber can connect a PagerDuty service, an Opsgenie account or team, or a Grafana IRM/OnCall stack. A single workflow then resolves the matching destination for each subscriber at delivery time.

    This supports products whose customers need alerts delivered to their own on-call systems. Your application collects the destination credential or webhook URL and registers it through the channel endpoints API. You do not need a separate provider integration or routing implementation for every customer.

    If a subscriber has no endpoint for the selected provider, Novu marks the Tool step as skipped for that subscriber. Other subscribers in the same trigger are unaffected.

    Send to shared or customer-defined HTTP endpoints

    Tool Webhook covers destinations that do not have a native Tool provider. It supports two routing modes:

    • Static: Configure one shared endpoint on the integration. Every delivery goes to that URL.
    • Dynamic: Register one or more tool_webhook endpoints for each subscriber. Novu sends the Tool step to every endpoint registered for that subscriber.

    Dynamic endpoints can define their own URL, headers, and POST, PUT, or PATCH method. This gives each subscriber control over where their notifications go without hardcoding those destinations in your workflow.

    Shape, sign, and inspect webhook deliveries

    Tool Webhook lets you configure the outbound JSON without replacing the content authored in the workflow. Novu builds the request from the integration-level JSON body, the rendered Tool step content, and step overrides. The rendered step content is always sent as content and wins if the integration body defines the same key.

    You can set a signing secret to add an HMAC-SHA256 digest in the X-Novu-Signature header. The receiving server can verify the signature against the raw request body before processing the delivery.

    Novu handles delivery retries and records the result in Activity, so you can inspect the workflow run, delivery status, and attempts without maintaining a separate webhook delivery system.

    To use the new channel:

    1. In the Integrations Store, select Connect Provider, open the Tool tab, and choose a provider.
    2. Add a Tool step to a workflow and write its content with the same Liquid variables available to other workflow steps.
    3. For subscriber-specific routing, register each destination through the channel endpoints API from your backend.
    4. Trigger the workflow as usual. Novu resolves the destination, delivers the Tool step, and records the result in Activity.

    Read the setup guides for PagerDuty, Opsgenie, Grafana, and Tool Webhook.

  • LangChain adapter for Novu Connect

    Connect a LangChain or LangGraph agent to Slack, Microsoft Teams, WhatsApp, Telegram, and email, with mapped conversation history and in-channel tool approval on the adapter-managed path.

    Langchain adapter for Novu Connect

    The LangChain adapter is now available through @novu/framework/langchain. It gives teams already using LangChain or LangGraph a direct path to Novu Connect without rebuilding their agent for each communication channel.

    Your agent or graph continues to run in your application. Novu Connect handles inbound channel events, conversation context, and delivery of the response back to the user.

    For OpenAI:

    npm install @novu/framework langchain @langchain/core @langchain/openai

    For Anthropic:

    npm install @novu/framework langchain @langchain/core @langchain/anthropic
    Good to know

    This is not limited to just OpenAI and Anthropic, you can use other LangChain provider keys.

    Return a config or invoke your own agent

    The adapter supports two handoff patterns based on how much of your LangChain setup you want it to manage.

    When you return a LangChainAgentConfig from onMessage, the adapter calls createAgent().invoke() in your application, maps ctx.history, and delivers the final assistant response:

    import { agent } from '@novu/framework/langchain';
    
    export const supportBot = agent('support-bot', {
      onMessage: async (_message, ctx) => ({
        model: 'openai:gpt-4o',
        system: 'You are a helpful support agent.',
      }),
    });

    If you already invoke a LangChain agent or LangGraph graph yourself, use toLangChainMessages(ctx.history), run your existing invoke() call, and return { messages }. Novu delivers the final assistant message. Tool approval is not managed by the adapter on this bring-your-own invocation path.

    Ask for approval in the channel

    On the config path, needsApproval lets you gate sensitive tools without building a separate approval flow for every channel. When the model calls a gated tool, Novu posts an Approve / Deny card and pauses the turn. After the user responds, the adapter replays the approval cycle from conversation history and continues the agent run.

    import { tool } from '@langchain/core/tools';
    import { agent } from '@novu/framework/langchain';
    import { z } from 'zod';
    
    const issueRefund = tool(
      async ({ orderId }) => ({ orderId, status: 'refunded' }),
      {
        name: 'issueRefund',
        description: 'Issue a refund for an order',
        schema: z.object({ orderId: z.string() }),
      },
    );
    
    export const supportBot = agent('support-bot', {
      onMessage: async () => ({
        model: 'openai:gpt-4o',
        system: 'You are a helpful support agent.',
        tools: [issueRefund],
        needsApproval: (toolCall) => toolCall.name === 'issueRefund',
      }),
    });

    This approval flow does not require a separate LangGraph checkpointer. The conversation history holds the information the adapter needs to resume the turn.

    Get started with npx novu connect --runtime langchain, then follow the LangChain quickstart. See the LangChain reference for config returns, custom invocation, approval gating, Next.js setup, and error handling.

Free to start, ready to scale

10K events/month free forever. From weekend projects to enterprise scale, we've got you covered.