Mindket
English
English简体中文
Sign inSign up
StoreMCPSkillsCalculatorLearnPriceBlog
Discover
Research312Writing59Business95Legal13Finance & accounting33Marketing & SEO175Developer1,422Productivity391Education81Companions14
Help and support
Language
English简体中文

Telegram bot that describes any photo you send with OpenAI

by Mindketn8n workflowProductivity
Newversion 1 · updated
Nodes6Apps2Like

Overview

  1. Get the Image
  2. Switch ( image or not )
  3. Analyze image
  4. Wait
  5. Send Content for the Analyzed image
  6. Update Telegram Error Message
Telegram bot that describes any photo you send with OpenAI, picture 1
‹2 / 2›

An n8n workflow that answers photos sent to a Telegram bot. The photo goes to an OpenAI vision model and the bot replies with a description. Any message that is not a photo gets a short reply asking for an image.

Telegram bot that describes any photo you send with OpenAI is an n8n workflow in the Productivity category, listed by Mindket. It is free. This is version 1, last updated on . Download it on this page or copy its JSON, then import it into an n8n canvas. Nothing runs until you activate it in n8n.

At a glance

Runs in
  • n8n Cloud: import the file, publish it and send your bot a photo
  • Your own n8n server: the same steps, with the server open to Telegram over HTTPS
  • Another question: change the Text Input field of Analyze image to ask for something other than a description
  • Another model: pick any OpenAI model that takes images in Analyze image

Platform

n8nWorkflow automation, in n8n’s cloud or on your own server
Connects 2 apps

How it works

Trigger

Telegram message

It starts when someone sends your bot a message. If the message holds a photo, n8n downloads it, and the trigger reads:

  • The photo, in its large size
  • The chat the message came from

Actions

  1. The workflow checks whether the message holds a photo
  2. A photo goes to an OpenAI vision model together with a question about what is in the image
  3. The model’s description is sent back to the chat
  4. Any other message waits three seconds and then gets a reply that asks for an image

Result

A Telegram reply that describes the photo you sent, or a short request for an image

Integrations

What you need before you start.

Telegram

Chat

Receives the photo and sends the description or the request for an image.

NeedsTelegram bot token

OpenAI

Vision model

Looks at the photo and writes the description.

NeedsOpenAI API key with access to a model that takes images

Workflow check

Checked by Mindket on 13 Nov 2026

Steps
6 steps, starting on a Telegram event
AI
1 step calls an AI model: Analyze image
Sign-ins
Signs in to: OpenAI and Telegram
API keys
No API key is left in the file
Step types
Mindket’s checker reads all 5 of its step types

Read from the workflow’s file, not from a run.

About

What it is

This n8n workflow describes photos for you inside Telegram. You send a photo to your bot, n8n downloads it and an OpenAI model looks at it. The bot replies in the same chat with a text description. Any other message waits three seconds and then gets a reply that asks for an image. Eight nodes do the work, two of them sticky notes.

Who it suits

People who want a quick description of a picture without leaving Telegram. It is made for people who already use n8n and have an OpenAI key. Skip it if you need to keep a record, since the workflow stores nothing.

Who uses it, and how
IndustryWho uses itHow they use itExample
AccessibilityPeople with low visionHear what a picture showsA user sends a photo of a street sign and reads what the model says it shows.
ContentSocial media managersDraft a caption or alt text from a pictureA manager sends a product photo and gets a description to edit.
RetailShop staffFind out what an item isA clerk sends a photo of an unlabelled box and asks what is in it.
EducationTeachersExplain a diagram or a pictureA teacher sends a photo of a diagram and reads the model’s description.

You need a little n8n: import a file, add two credentials and make a Telegram bot.

Where to use it

Any n8n that Telegram can reach over HTTPS and that can reach OpenAI will run it. The bot answers in the chat the photo came from.

  • n8n Cloud: import the file, publish it and send your bot a photo
  • Your own n8n server: the same steps, with the server open to Telegram over HTTPS
  • Another question: change the Text Input field of Analyze image to ask for something other than a description
  • Another model: pick any OpenAI model that takes images in Analyze image

Setup

Download saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Setup is a Telegram bot and an OpenAI key.

What you need first
NeedDetail
n8nOn n8n Cloud or on a server of your own, open to Telegram over HTTPS
TelegramA bot and its token from BotFather
OpenAIAn API key with access to a model that takes images
  1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
  2. Add a Telegram credential with your bot’s token and select it in Get the Image, Send Content for the Analyzed image and Update Telegram Error Message.
  3. Add an OpenAI credential and select it in Analyze image. The node is set to gpt-6-luna, so check Good to know before you rely on it.
  4. Press Execute workflow, send your bot a photo and check the reply.
  5. Publish the workflow so the bot answers all day.

How to use it

Send your bot a photo as a photo, not as a file. The reply comes after a few seconds and says what the model sees. Any text or file gets a reply that asks for an image.

What to change for a different result
To changeWhere
The question asked about the photoThe Text Input field of Analyze image
How long the answer can beThe Length of Description option in Analyze image
The modelThe Model field of Analyze image
The reply to a message that is not a photoThe text of Update Telegram Error Message
How long it waits before that replyThe amount in Wait

Example

You send a photo of a desk with a laptop and a plant on it.

The bot replies with a few sentences that name the laptop and the plant and say where they sit. Treat the description as a model’s guess, since it can miss or misread details.

For newcomers and experts

If you are new. n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. A Telegram bot is an account that a program controls. Telegram hands each message sent to it over to n8n. A vision model takes an image and a question and writes an answer.

For experts. Get the Image downloads a photo in its large size into the binary field data. Switch ( image or not ) sends a message that has a photo field to Analyze image. Every other message goes to Wait. Analyze image is the OpenAI node’s Analyze Image operation with a binary input and Simplify on, so it returns one content field. That field is what Send Content for the Analyzed image sends.

Where to look closer
NodeWhat it doesWhat to know
Get the ImageTakes each new message and downloads its photoIt listens for message updates only
Switch ( image or not )Checks the message for a photo fieldA photo sent as a file has none and takes the other path
Analyze imageSends the photo and a question to an OpenAI modelThe question is the node’s default, which asks what is in the image
Send Content for the Analyzed imageSends the description back to the chatIt turns off n8n’s attribution line
WaitPauses three secondsIt only delays the reply to a message that is not a photo
Update Telegram Error MessageAsks the sender for an imageIts text is Please Upload an Image

Good to know

The file called a model that OpenAI shut down in 2024, and nothing checks who sends the bot a photo.

  • Model: the file’s Analyze image node was version 1.3. n8n’s source shows that version sending gpt-4-vision-preview, and OpenAI’s deprecations page, read on 12 November 2026, lists that model as shut down on 6 December 2024. Version 1.4 and later let you choose the model, so Mindket’s copy uses 1.4 with gpt-6-luna, the replacement OpenAI names for gpt-4o. Its model page lists text and image input. The node’s own default, gpt-4o, is listed for shutdown on 14 January 2027
  • Answer length: the node caps the answer at 300 tokens unless told otherwise. gpt-6-luna reasons before it answers, and OpenAI’s reasoning guide says the cap counts reasoning tokens and can run out before any text appears. So Mindket’s copy sets 1500. This was not run from here, so send a photo and check that a description comes back
  • Photos only: a photo sent as a file has no photo field, and a text message has none either, so both get the reply that asks for an image
  • Open bot: nothing checks who sends a photo, so anyone who finds the bot can spend your OpenAI credit. Telegram Trigger version 1.2 and later offers Restrict to User IDs under Additional Fields, and the file’s trigger is version 1.1. To limit it, swap in a current Telegram Trigger, turn on Download Images/Files and fill in Restrict to User IDs
  • Your data: each photo goes to OpenAI
  • Not run: nothing here was run against Telegram or OpenAI. The notes come from n8n’s source and the vendors’ pages
  • Credentials: the JSON holds none

README and files

README

Telegram bot that describes any photo you send with OpenAI

An n8n workflow that answers photos sent to a Telegram bot. The photo goes to an OpenAI vision model, and the bot replies with a
description. Any message that is not a photo gets a short reply that asks for an image.

What is inside

  • workflow.json: the workflow as n8n exports it, 8 nodes, two of them sticky notes. It holds no credential. One note holds
    Mindket's setup notes, and the other is a coloured frame behind the nodes.
  • README.md: these notes.

What we changed from the original copy

  • The Telegram nodes and the OpenAI node were linked to credentials in their author's n8n. The links are gone, so select your
    own.
  • Analyze image was version 1.3 of the OpenAI node, which sends the model gpt-4-vision-preview. OpenAI shut that model down
    in 2024. The node is now version 1.4, with the model set to gpt-6-luna and the answer length raised to 1500 tokens.
  • The long sticky note was rewritten. It named the author's own bot, and it described two nodes, a Merge and an Aggregate, that
    are not in the file. The new note has the setup steps and a few warnings.
  • The Telegram Trigger's and the Wait node's webhook ids are gone, and n8n makes new ones when you import the file.
  • The export's instance id is gone.
  • Nothing else.

Before you start

  • An n8n instance that Telegram can reach over HTTPS, on n8n Cloud or on your own server.
  • A Telegram bot and its token, made by messaging BotFather in Telegram.
  • An OpenAI API key with access to a model that takes images.

Set it up

  1. In n8n, create a workflow, open its menu and choose Import from File. Pick workflow.json.
  2. Add a Telegram credential with your bot's token. Select it in Get the Image, Send Content for the Analyzed image and
    Update Telegram Error Message.
  3. Add an OpenAI credential and select it in Analyze image. The node is set to gpt-6-luna, so read Good to know before you
    rely on it.
  4. Press Execute workflow, send your bot a photo and check the reply.
  5. Publish the workflow so the bot answers all day.

How a run goes

  1. Get the Image takes a new message. It listens for message updates only. When the message holds a photo, the node
    downloads it in its large size into the binary field data.
  2. Switch ( image or not ) checks the message for a photo field. A message that has one goes to Analyze image. Every other
    message goes to Wait.
  3. Analyze image sends the photo and the node's default question, which asks what is in the image, to the model. With Simplify
    on, it returns one content field.
  4. Send Content for the Analyzed image sends that text to the chat the photo came from, with n8n's attribution line turned off.
  5. Wait pauses three seconds, and Update Telegram Error Message then sends the reply Please Upload an Image .....

Change the look

  • The question asked about the photo: the Text Input field of Analyze image.
  • How long the answer can be: the Length of Description option in Analyze image.
  • The model: the Model field of Analyze image.
  • The reply to a message that is not a photo: the text of Update Telegram Error Message.
  • How long it waits before that reply: the amount in Wait.

Good to know

  • Model: n8n's source shows version 1.3 of the OpenAI node's Analyze Image sending gpt-4-vision-preview, and OpenAI's
    deprecations page, read on 12 November 2026, lists that model as shut down on 6 December 2024. Version 1.4 and later let you
    choose the model. Mindket's copy uses 1.4 with gpt-6-luna, the replacement OpenAI names for gpt-4o. The model page lists
    text and image input for it. The node's own default, gpt-4o, is listed for shutdown on 14 January 2027.
  • Answer length: the node caps the answer at 300 tokens unless told otherwise. gpt-6-luna reasons before it answers. OpenAI's
    reasoning guide says the output cap counts reasoning tokens and can run out before any text appears. So Mindket's copy sets
    1500. This was not run from here, so send a photo and check that a description comes back.
  • A photo sent as a file has no photo field, and neither has a text message, so both get the request for an image.
  • Nothing checks who sends the bot a photo, so anyone who finds the bot can spend your OpenAI credit. Telegram Trigger version 1.2
    and later offers Restrict to User IDs under Additional Fields, and the file's trigger is version 1.1. To limit it, swap in a
    current Telegram Trigger, turn on Download Images/Files and fill in Restrict to User IDs.
  • Your data: each photo goes to OpenAI.
  • Not run: nothing here was run against Telegram or OpenAI. These notes come from n8n's source and the vendors' pages, read on 12
    November 2026.
Files
NOTICE.md  631 B
README.md  5 KB
workflow.json  6 KB

3 files, 12 KB, downloaded as a zip.

What’s new

  • Version 1 · 2026-11-12

Reviews

No ratings yet

Write a review

No reviews yet.

More from Mindket

See all from Mindket

anaksor-ocr

Sokheng · Kit

Reads Khmer and English text from documents, receipts and photos of signs, using a detector for text lines and a recogniser for each line. Suits developers who…

New

Qwen2.5-7B-Instruct-GGUF by Qwen

Qwen · Kit

Qwen2.5-7B-Instruct as GGUF files in eight quantization sizes, to chat with locally through llama.cpp. For people who want a capable 7.6B model on their own co…

New

ACE-Step 1.5 XL SFT

ACE-Step · Kit

Generates music from text with a 4B diffusion decoder, 50 steps and guidance control, aiming for higher audio quality than the base. For creators with a GPU of…

New

Similar in Productivity

See all in Productivity

Guizang PPT Skill

op7418 · Skill

Skill that makes single-file HTML slide decks, slide illustrations and platform covers in two visual styles, with a presenter mode for rehearsal. For speakers…

New29.8k

Frontend Slides

zarazhangrui · Skill

Builds HTML presentations from scratch or from a PowerPoint file and shows visual style previews to pick from, for people who want web slides with no CSS or Ja…

New32.3k

IPTV

iptv-org · Kit

Collects publicly available live TV channels from many countries into playlists that any video player with live streaming can open. Made for viewers and for de…

New144k