You do not need a separate app or a special command to generate images in Telegram with @DoublegramAIBot. Describe what you want in normal language in your private chat with the bot, and it creates the image and sends it back — usually together with a short text reply.
This guide covers how to ask for images, how to pick the right image model, and how to write prompts that get better results. Start from @DoublegramAIBot in a private chat.
How image generation works
When you ask the bot to create or generate an image, it:
- Understands your request from your message.
- Builds a detailed visual description behind the scenes.
- Generates the image using your selected image model.
- Sends the picture to you in Telegram (1024×1024 pixels).
You can also ask for an image as part of a longer request — for example: Write a short social post about summer travel and generate a matching image.
Try it: Send /model → Image Model → pick a model → then type: Generate an image of a cozy coffee shop on a rainy evening, warm lighting, watercolor style.
Choose your image model first
Image generation uses a separate setting from your chat model. Set it once with /model:
- Send
/model(or Settings → Models). - Tap Image Model (not Text Model).
- Pick a provider — OpenAI or Google Gemini.
- Tap the model you want. Each button shows the cost per image (for example 4/img).
Available image models:
- GPT Image 1.5 — OpenAI's latest image model. Strong all-round quality.
- GPT Image 1 — OpenAI image generation. Higher credit cost per image.
- Imagen 3 — Google's image model. Default for new users; good balance of quality and cost.
Send /status to see your current text model; the image model is managed through /model → Image Model.
How to ask for an image
Use clear, natural language. Phrases that work well:
- Generate an image of…
- Create a picture of…
- Make an illustration of…
- Draw a…
Be specific. The more detail you give, the closer the result matches what you imagine:
- Subject — what is in the scene?
- Style — photorealistic, cartoon, watercolor, minimalist, 3D render…
- Mood and lighting — bright, dark, neon, golden hour…
- Colors — if they matter to you, say them.
Example prompts:
- Generate an image of a golden retriever wearing sunglasses on a beach, cartoon style, bright colors.
- Create a minimalist logo for a coffee brand — black and white, clean lines.
- Make a photorealistic image of a modern kitchen with marble counters and morning sunlight.
Generate vs analyze — don't mix them up
Doublegram AI handles images in two different ways:
- Generate a new image — You describe what you want and the bot creates it. Use phrases like generate, create, or draw.
- Analyze an image you sent — You send a photo and ask questions about it (Vision). Use phrases like what is in this photo? or describe this image.
If you send a photo in Telegram and say use this image in my post, the bot uses your photo — it does not generate a new one. To get a brand-new picture, ask explicitly to generate or create an image.
Tip: Want a post with a fresh visual? Say Create a post about fitness with a generated image of someone running in a park at sunrise. One message, text + image.
Credits and costs
Generating an image costs credits in addition to your normal chat response. The image cost depends on the image model you selected — shown on each button in /model → Image Model.
- Image generation is charged only when the image is created successfully.
- Check your balance with
/creditsbefore generating several images in a row. - If you run out of credits, the bot may reply in text but skip the image — top up via
/credits→ Get AI Credits.
Quick reference
- Set image model:
/model→ Image Model - Ask naturally: Generate an image of…
- Default image model for new users: Imagen 3
- Output size: 1024×1024
- Check balance:
/credits