> For the complete documentation index, see [llms.txt](https://dnotifier.gitbook.io/product-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://dnotifier.gitbook.io/product-docs/platform-overview/overview.md).

# Build AI agents that work together

DNotifier is a developer platform for shipping AI-powered, realtime applications without stitching together half a dozen services. You get **AI agents**, **multi-agent orchestration**, **AI workflows**, **AI infrastructure**, **observability**, and **realtime communication** — all through official SDKs and a unified dashboard at [app.dnotifier.com](https://app.dnotifier.com).

{% hint style="success" %}
**Start here** if you want the big picture. When you are ready to build, follow [Getting set up](/product-docs/getting-started/getting-set-up.md) or jump to [Explore the product](/product-docs/platform-overview/explore-the-product.md).
{% endhint %}

***

## What DNotifier does

DNotifier helps teams connect users, services, and AI in production. Instead of operating separate messaging infrastructure, LLM gateways, RAG pipelines, workflow engines, and monitoring stacks, you integrate one platform and focus on your product logic.

Every app you create in the dashboard gets credentials, plan limits, and access to the capabilities below. Your client or server connects with an SDK, authenticates with your app secret, and starts sending realtime messages, running AI prompts, or executing multi-step workflows.

***

## Six platform pillars

### 1. AI Agents

Define **named agents** — reusable units of AI logic with a clear purpose. An agent receives input, calls AI or your knowledge base, and returns structured output. Agents are the building blocks for assistants, support bots, classifiers, and content generators.

Use agents standalone (single prompt → response) or compose them inside workflows. The SDK exposes `defineAgent` so you declare behavior in code and keep prompts, tools, and business rules in one place.

**Typical uses:** support agents, knowledge bots, intent classifiers, summarizers, internal copilots.

→ Learn more: [AI use cases](/product-docs/platform-overview/use-cases/ai.md) · [Define an agent](/product-docs/workflows-and-agents/define-agent.md)

***

### 2. Multi-Agent Orchestration

When one agent is not enough, DNotifier lets you **chain and route between multiple agents**. A router agent can classify intent; specialized agents handle search, billing, or general Q\&A. Shared workflow state passes context from step to step so each agent sees what came before.

Multi-agent orchestration is deterministic at the workflow level: you control branching, fallbacks, and which agent runs next — not a black-box autopilot.

**Typical uses:** intent routing, specialist handoffs, pipeline stages (research → draft → review).

→ Learn more: [Example: Multi-agent pipeline](/product-docs/workflows-and-agents/example-multi-agent.md)

***

### 3. AI Workflows

**Workflows** are multi-step pipelines with an entry point, shared state, and optional observability. A workflow can run agents sequentially, call the knowledge base, branch on results, and return a final payload to your app.

Workflows are defined in code (`Workflow` + `registerAgents`) and executed through the SDK. Enable observability to record executions and steps in the DNotifier dashboard — useful for debugging production AI behavior.

**Typical uses:** RAG Q\&A pipelines, content generation, approval flows, automated support triage.

→ Learn more: [Workflows overview](/product-docs/workflows-and-agents/overview.md) · [Build a workflow](/product-docs/workflows-and-agents/build-workflow.md)

***

### 4. AI Infrastructure

DNotifier provides the **runtime and APIs** your AI features need — without you operating vector databases, session stores, or LLM proxy layers yourself.

| Capability               | What you get                                                                  |
| ------------------------ | ----------------------------------------------------------------------------- |
| **Prompts & sessions**   | `sendAI`, session IDs, multi-turn conversations                               |
| **Knowledge base (RAG)** | Add, update, list, delete documents; semantic search                          |
| **Chat-style messages**  | System/user/assistant message arrays with optional RAG                        |
| **History**              | Fetch and delete AI conversation history per user                             |
| **Transport**            | HTTP for request/response AI; WebSocket when you need a persistent connection |

Plan limits (AI requests per month, knowledge base word count, etc.) are returned at connect time via `getPlanLimits()` so your app can adapt UI and behavior.

**Typical uses:** in-app assistants, documentation Q\&A, semantic search over product data.

→ Learn more: [AI overview](/product-docs/ai/overview.md) · [Knowledge base](/product-docs/ai/knowledge-base-overview.md)

***

### 5. Observability

Ship AI and workflows with visibility into what happened in production.

| Surface                    | What it tracks                                                                                          |
| -------------------------- | ------------------------------------------------------------------------------------------------------- |
| **Session logging**        | Enable `logs: true` on the SDK to report AI session start, completion, and token usage to the dashboard |
| **Workflow observability** | Set `observability: true` on a workflow to record execution and step telemetry in the dashboard         |
| **Plan usage**             | Dashboard and auth response reflect messages, AI requests, and knowledge base limits                    |

Observability is built into the platform — not a bolt-on APM integration you configure separately for every AI call.

→ Learn more: [Session logging](/product-docs/ai/session-logging.md) · [Workflow observability](/product-docs/workflows-and-agents/observability.md)

***

### 6. Realtime Communication

DNotifier delivers **realtime 1:1 messaging** between users, devices, and backend services over **WebSocket** (with HTTP send for server-side delivery). Messages are addressed by **sender ID** and **receiver ID** — direct, explicit routing without broadcast topics.

| Feature                        | Description                                                   |
| ------------------------------ | ------------------------------------------------------------- |
| **Text & structured payloads** | JSON messages with types like `text`, `image`, `audio`, `doc` |
| **Binary messaging**           | Raw buffers for custom protocols                              |
| **Large payloads**             | Automatic chunking and reassembly                             |
| **Multiple receivers**         | Send the same payload to several receiver IDs in one call     |
| **Server-side send**           | HTTP RPC to deliver messages to connected clients             |

Realtime communication is the **transport layer**. **Chat** is a use case built on top (history, persistence, UI patterns) — see [Chat overview](/product-docs/chat/overview.md).

**Typical uses:** live dashboards, in-app notifications, collaborative UI sync, backend-to-client updates.

→ Learn more: [Realtime communication overview](/product-docs/realtime-communication/overview.md) · [Realtime use cases](/product-docs/platform-overview/use-cases/realtime-communication.md)

***

## How the pillars fit together

```
┌─────────────────────────────────────────────────────────────┐
│                     Your application                        │
│  (web, mobile, server, microservices)                       │
└──────────────────────────┬──────────────────────────────────┘
                           │ Official SDKs (npm, pub.dev)
                           ▼
┌─────────────────────────────────────────────────────────────┐
│                      DNotifier platform                     │
│  ┌──────────────┐  ┌──────────────┐  ┌──────────────────┐ │
│  │   Realtime   │  │  AI + RAG    │  │    Workflows     │ │
│  │  1:1 msgs    │  │  sessions    │  │  multi-agent     │ │
│  └──────────────┘  └──────────────┘  └──────────────────┘ │
│  ┌────────────────────────────────────────────────────────┐ │
│  │              Observability & dashboard                 │ │
│  └────────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────┘
```

Many products use **realtime communication** for live UX, **chat** for conversations, **AI** for intelligence, and **workflows** for structured automation — often in the same app. See [Combined use cases](/product-docs/platform-overview/use-cases/combined.md).

***

## Who should use DNotifier?

| Role                    | Why DNotifier                                                                      |
| ----------------------- | ---------------------------------------------------------------------------------- |
| **Founders**            | Ship chat, AI assistants, or live features without hiring infra specialists        |
| **Frontend developers** | WebSocket clients, chat UIs, live progress, AI panels in React, Vue, Flutter, etc. |
| **Backend developers**  | Server-side message delivery, AI/RAG APIs, workflow runners, webhook bridges       |
| **Full-stack teams**    | One platform, one dashboard, consistent SDKs across client and server              |

***

## Next steps

| Goal                                 | Start here                                                                                                                      |
| ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------- |
| See product capabilities at a glance | [Explore the product](/product-docs/platform-overview/explore-the-product.md)                                                   |
| Understand SDKs, limits, and pricing | [What you get](/product-docs/platform-overview/what-you-get.md) · [Pricing & plans](/product-docs/platform-overview/pricing.md) |
| Pick a scenario                      | [Use cases](/product-docs/platform-overview/use-cases.md)                                                                       |
| Plan your integration                | [Developer integration guide](/product-docs/platform-overview/developer-integration-guide.md)                                   |
| Create an account                    | [Register your account](/product-docs/getting-started/register-account.md)                                                      |
| Install an SDK                       | [Install an SDK](/product-docs/getting-started/installation.md)                                                                 |

{% hint style="info" %}
**Official links:** [dnotifier.com](https://www.dnotifier.com) · [app.dnotifier.com](https://app.dnotifier.com) · [Discover](https://www.dnotifier.com/discover) · [npm](https://www.npmjs.com/package/@dnotifier-realtime/dnotifier) · [pub.dev](https://pub.dev/packages/dnotifier)
{% endhint %}