# Build a comprehensive multimodal assistant on Telegram: n8n workflow

> Started by an incoming Telegram message, and works across Telegram and in-memory vectors. It puts OpenAI behind a retrieval step, so answers are drawn from...

Source: https://millionrounds.com/workflow/build-a-comprehensive-multimodal-assistant-on-telegram (Millionrounds; every figure on the page links to its filing or article)

[Automation library](https://millionrounds.com/workflows) · [Chat assistants](https://millionrounds.com/workflows?category=Chat+assistants#library)

# Build a comprehensive multimodal assistant on Telegram

Started by an incoming Telegram message, and works across Telegram and in-memory vectors. It puts OpenAI behind a retrieval step, so answers are drawn from documents you supply rather than from the model alone. 33 steps, 2 branches: large, and worth reading before you trust it with anything live.

Large Calls a model: OpenAI Retrieval

[Download free](https://millionrounds.com/login?next=%2Fapp%2Fworkflow%2Fbuild-a-comprehensive-multimodal-assistant-on-telegram)

**33 steps** Large: read it before trusting it with anything live

**2** Branches: places the flow splits

**0** Loops or waits

**2** Services it connects to

## The flow

Starts it 1 Model 17 Logic 10 Service 5

The steps in the order they run, from what starts it.

- 1 **Receive Message** Telegram trigger *Starts it*
- 2 **Window Buffer Memory** Memory Buffer Window *Model*
- 3 **Embeddings OpenAI1** Embeddings Open Ai *Model*
- 4 **Token Splitter** Text Splitter Token Splitter *Model*
- 5 **Image Generator** Tool Http Request *Model*
- 6 **Basic Google Search** Tool Serp Api *Model*
- 7 **Webpage Scraper** Tool Http Request *Model*
- 8 **Calculator** Tool Calculator *Model*
- 9 **OpenAI Chat Model** Lm Chat Open Ai *Model*
- 10 **Think** Tool Think *Model*
- 11 **Embeddings OpenAI** Embeddings Open Ai *Model*
- 12 **API Setup** Set *Logic*
- 13 **J.A.R.V.I.S.** Agent *Model*
- 14 **Simple Vector Store** Vector Store In Memory *Model*
- 15 **Default Data Loader** Document Default Data Loader *Model*
- 16 **Simple Vector Store1** Vector Store In Memory *Model*
- 17 **Switch** Switch *Logic*
- 18 **If Audio Response** If *Logic*
- 19 **Format-doc-output** Set *Logic*
- 20 **Download Image** Telegram *Service*
- 21 **Get Audio File** Telegram *Service*
- 22 **Format-text-output** Set *Logic*
- 23 **Download Document** Telegram *Service*
- 24 **Split MSG** Code *Logic*
- 25 **Generate Audio** Open Ai *Model*
- 26 **Fix File Extension** Code *Logic*
- 27 **Transcribe** Open Ai *Model*
- 28 **Convert Doc** Code *Logic*
- 29 **Text Response** Telegram *Service*
- 30 **Audio Response** Telegram *Service*
- 31 **Analyze Image** Open Ai *Model*
- 32 **Format-voice-output** Set *Logic*
- 33 **Format-image-output** Set *Logic*

## How it's built

- **Starts on:** An incoming Telegram message
- **Steps:** 33, with 2 branches
- **Size:** Large: read it before trusting it with anything live
- **Connects to:** Telegram, in-memory vectors
- **Model:** OpenAI
- **Retrieval:** Answers are drawn from documents you supply, not from the model alone

## Questions

What does it need to run?

An n8n account, plus your own Telegram and in-memory vectors accounts, and a key for OpenAI. You connect your own accounts after importing it; none are included.

What starts it?

It starts on an incoming Telegram message.

Does it use AI?

Yes. It calls OpenAI. Its answers draw on documents you give it, not on the model alone. Model calls are billed by the provider, not by n8n.

How long does it take to set up?

Take your time. At 33 steps it is large: read it before trusting it with anything live.

How do I import it?

Copy the workflow, open a new workflow in n8n and paste. The steps appear on the canvas, ready for your own accounts.

**Before you run it**

Nothing here has been tested by us. It is somebody's working automation, read and described from its own steps. Treat the connections it wants as a list of accounts it will touch, and read the 33 steps before you point it at anything that matters.

## Run it in n8n

- **Create your n8n account** Free to start. n8n is where the workflow runs. [Create an n8n account ↗](https://millionrounds.com/n8n)
- **Download the workflow** Free with a Millionrounds account. [Download free](https://millionrounds.com/login?next=%2Fapp%2Fworkflow%2Fbuild-a-comprehensive-multimodal-assistant-on-telegram)
- **Paste it into n8n** Open a new workflow in n8n and paste. Then connect your own accounts.

Stored with the author's traces removed: canvas notes, instance identifiers, saved credential names and any live document they had left behind.

## Connects to

Each opens the other workflows that use it.

[Telegram](https://millionrounds.com/workflows?service=Telegram#library)[in-memory vectors](https://millionrounds.com/workflows?service=in-memory+vectors#library)

## Workflows like it

Also chat assistants, with the same services where possible, closest in size first.

- [**Plan daily restaurant prep lists with DeepSeek** 33 steps · calls a model · Gmail, Google Sheets](https://millionrounds.com/workflow/plan-daily-restaurant-prep-lists-with-deepseek)
- [**Voice-to-task manager with Telegram and OpenAI** 33 steps · calls a model · Notion, Telegram](https://millionrounds.com/workflow/voice-to-task-manager-with-telegram-and-openai)
- [**Assist hospital doctors with Telegram and OpenAI** 34 steps · calls a model · Gmail, Google Sheets](https://millionrounds.com/workflow/assist-hospital-doctors-with-telegram-and-openai)
- [**Classify and auto-reply to Gmail with OpenAI** 34 steps · calls a model · Gmail, Google Sheets](https://millionrounds.com/workflow/classify-and-auto-reply-to-gmail-with-openai)
- [**Email management system with OpenAI and Pinecone DB** 34 steps · calls a model · Gmail, Pinecone](https://millionrounds.com/workflow/email-management-system-with-openai-and-pinecone-db)
- [**Evaluate job fit and generate application assets** 34 steps · calls a model · Apify, Gmail](https://millionrounds.com/workflow/evaluate-job-fit-and-generate-application-assets)