Unlimited Generations.
Both include unlimited generations. Zero per-image fees.Both include unlimited generations. Zero per-image fees.
$199/mo
$449/mo
Custom Model For Your Character
Generate memes in telegram with a custom model that is trained for free and delivered in 24 hours.Generate memes in telegram with a custom model. Trained free and delivered in 24 hours.
Upload
8-20 character images via dashboard.
Train
Custom model fine-tuned on dedicated GPU.
Deploy
Deployed to private test group for review.
Generate
Users prompt, bot replies with images.
Good Training Images
- High resolution (1024x1024+)
- Multiple angles and poses
- Consistent character design
- Clean backgrounds when possible
- Full body and close-up shots
Avoid These
- Low resolution or blurry images
- Heavy watermarks or overlays
- Inconsistent character details
- Multiple characters in one image
- Extreme cropping or partial views
Revisions Included
Not satisfied with the initial training? We'll adjust until it's right.
Trigger Word + Prompt
Community member types trigger followed by description. Bot queues, generates, delivers. Works in any group.
Shared or Dedicated
Both tiers deliver unlimited, high-quality generations. Choose based on speed and multi-character needs.
Shared Compute
- Generation Time: ~15 seconds
- Queue: Round-robin (up to 4 projects)
- Characters: 1
- Monthly Capacity: Unlimited
- Watermark: Default (removable)
- Training: $0 included
- Revisions: 2 free rounds
Best for: Testing viability, smaller communities
Dedicated Compute
- Generation Time: 4 seconds guaranteed
- Queue: None (exclusive GPU)
- Characters: Up to 8 (free, 11s gen)
- Monthly Capacity: Unlimited
- Watermark: None
- Custom Resolutions: Available
- LoRA Hotswapping: Dynamic per-trigger
Best for: Established projects, high-volume communities
Dedicated Compute. Zero Cloud Dependencies.
We don't rent cloud GPUs. We own the servers outright. Air-gapped, encrypted tunnels, no AWS/Azure/GCP in the chain.
Your Community
Telegram group members send prompts
Telegram Webhook
Receives and validates requests
GPU Server
Air-gapped, owned hardware
All ports closed
Air-Gapped
GPU servers isolated from public networks
Encrypted Tunnels
Cryptographically protected connections
DDoS Resistant
Zero attack surface from open endpoints
Owned Hardware
Direct control, no cloud dependencies
We don't depend on
When major cloud providers go down, we stay up. Your bot keeps generating.
Shared Compute: Round RobinPrioritized Round Robin
Our Shared Compute plan for Telegram bots uses a round-robin queue system that rotates one image at a time between projects.
Project A has 10 prompts. Project B has 5 prompts. Your project sends 1 prompt. The system executes one generation for A, then one for B, then yours—it doesn't wait for A's entire queue to clear. Everyone advances together.
You're Already Halfway There.
Same training. Same character. Add Twitter/X and save $400.
Shared vs Dedicated ComputeShared vs Dedicated
Transparent overview of both telegram packages.
Throttle or Block Malicious Users
Limit or block any users that break your rules.
1 user sent 200 requests — everyone else waits in line
No single user can slow down the group
DAILYLIMIT
Configurable generations per day
BLOCK
Complete access revocation
EDIT HISTORY
Full audit trail of changes
How Group Authorization Works
Every bot starts locked — no chat can trigger image generation until you authorize it. To add a group, invite the bot and have anyone type /getid. The bot records the group's name, type, and member count, and it appears in your dashboard's Groups tab. Select it, click Authorize, and the bot starts responding there within seconds. Revoke access the same way — one click, immediate effect.
Add Bot to Chat
Invite the bot to any Telegram group or DM
Send /getid
Bot records the chat's ID, name, and member count
View in Dashboard
The discovered chat appears in your Groups tab
Approve the Chat
Click to authorize — the chat can now use the bot
Generation Active
Members of that chat can generate images
Your Bot, Your Commands
Add custom trigger/response pairs, project links, docs, anything. Forward slash or not, both work. Toggle on and off anytime.
Add Commands To Your Bot
Define trigger words and the bot replies with your text. No slash needed — “ca” works the same as “/ca”.
Add Unlimited Commands
Add unlimited commands to your bot that function in all groups, such as ca or chart, forward slash isn't needed.
Edit the Trigger or Response
Change the trigger word or the response text directly from the dashboard. No need to delete and re-create.
Owner Gets a Confirmation
When a command is added, edited, or toggled, you receive a Telegram DM with the details of what changed.
Applied on Save
When you save a change, the bot's command file is updated and the service restarts. The new command is live within seconds.
Get Notified the Moment Anything Changes
When you authorize a new group, set a rate limit, or add a command the bot sends you a message on Telegram immediately.
When You Authorize a Group
Authorized: "Crypto Alpha" Type: supergroup Members: 1,500
When You Set a Rate Limit
User: @spammer_420 Type: DAILYLIMIT Limit: 50/day
When You Add a Command
Trigger: /chart Response: dexscreener.com/... Status: Active
How To Enable Notifications
After purchase, you'll be walked through a one-time notification setup using the Telegram username you entered during bot creation.
Only the Telegram username you provided during bot creation has access to management controls. That means viewing who is rate-limited or blocked, enabling and receiving owner notifications, and authorizing which groups can use the bot. Group members and other admins cannot see or modify these settings — they are locked to your account.
Send /start to Your Bot
Open a DM with the bot and send /start — this is a one-time step
Bot Records Your Chat ID
The bot saves your Telegram ID so it knows where to send messages
Notifications Active
Every change made to your bot triggers a DM sent to your Telegram
Dedicated Plans Only
Eight Characters.
One Bot.
Dynamic Switching.
Each character gets its own trigger word. Community members choose which character to generate. LoRA weights hotswap dynamically per request. 11-second generation.
1 Included
Base Character
Free
Each Additional
8 Maximum
Total Characters
Zero
Cross-Contamination
Edit Any Image Instantly
Included In Dedicated Plans
They Charge Per Image. We Don't.
Other providers charge $0.05-$0.20 per generation. At 1,000 images/month, that's $50-$200 on top of your subscription. We charge $0.00 per image.
Zero Per-Image Fees
Competitors charge $0.05-$0.20 per generation. Generate 10,000 images with them, pay $500-$2,000 extra. Generate 10,000 images with us, pay $0 extra.
Our Hardware
We own the GPUs. No third-party API calls. No per-inference costs passed to you. That's why unlimited actually means unlimited.
Truly Unlimited Generation
Your GPU processes images sequentially, one after another. No monthly caps, no overage fees. It generates as fast as prompts come in.
Round-Robin Queue
Your 1 prompt doesn't wait behind someone's 100. System rotates: A gets 1, B gets 1, C gets 1, back to A. Maximum wait: 33 seconds.
$199 or $449. Period.
Training: $0. Deployment: $0. Support: $0. Per-image fees: $0. The price listed is the price you pay. Every month. No surprises.
Test Before You Commit
Evaluate in a private group. Not satisfied with model quality? Full refund. Two retraining rounds included at no cost.
Questions
Direct answers.
Other providers charge $0.05-$0.20 per generation on top of their subscription. At 5,000 images/month, that's $250-$1,000 extra. We charge $0.00 per image. $199 or $449 monthly is your total cost.
Yes. You get a physical GPU that processes images one after another. There are no monthly caps, no per-image fees, no overage charges. It generates as fast as prompts come in—the only limit is physics.
They use third-party cloud APIs (Replicate, RunPod, etc.) that charge per inference. They pass that cost to you. We own our GPUs—no per-inference cost, no markup to pass on.
$199: GPU shared with 3 others, ~15s generation, unlimited, round-robin queue (max 15s wait). $449: 100% dedicated GPU, 4s guaranteed, unlimited, multi-character support, no queue.
Under 24 hours. Usually same day. Upload 8-20 images, we train the model, configure the bot, set up webhooks. You test in a private group, approve, go live.
Test before committing. Evaluation in private group. Not satisfied? Full refund. After approval, you get 2 free retraining rounds with different images.
No. Others charge $199-$899 for training. We include it. The monthly price covers training, deployment, support, everything.
Unlimited groups. Same bot, no extra cost. Add it to 1 group or 100 groups.
24-hour grace period. Then bot goes offline. Model preserved indefinitely. Pay next month, bot reactivates instantly with all settings intact.
Yes. Changes take effect next billing cycle. Bot, model, settings all transfer. No retraining required.
Have an established project already? If your project is older than 30 days, reach out to us via DMs on X @memescaleai, and we can train a model for you, at zero cost, and provide three days of entirely free compute to test the model on Telegram.
Unlimited images. Pick your speed.
$449/mo
Dedicated · 5s guaranteed