Multi-agent Setup Telegram
Hello,
I am designing a multi-agent architecture in OpenClaw with multiple Telegram bots and need clarification on the correct configuration approach.
Current Setup
I have created four Telegram bots:
Larry
Bob
Sentinel
Alfred
My goal is the following:
Each bot should correspond to a dedicated agent.
Each agent should use a different model configuration.
Larry should act as the orchestrator.
Larry must be able to spawn and coordinate the other agents.
At the same time, I want to be able to directly interact with each bot independently in its own Telegram chat.
Desired Behavior
When I write to Larry's Telegram bot, Larry responds and can delegate tasks to Bob, Sentinel, or Alfred.
When I open Bob’s Telegram chat, I want to talk directly to Bob (without going through Larry).
Same for Sentinel and Alfred.
Each agent must remain isolated in its own Telegram context.
No message leakage between agents.
Model configurations must remain independent per agent.
Questions
What is the correct configuration pattern to:
Bind one Telegram bot per agent?
Ensure strict session isolation per bot?
Allow orchestration while still permitting direct conversations?
Should each agent define:
Its own Telegram transport?
Its own workspace?
Its own binding rules?
What is the recommended architecture for:
Orchestrator agent with delegation privileges
Specialist agents accessible both directly and via orchestration?
Is there a minimal working example of a configuration that supports:
Multiple bots
Multiple agents
Independent models
Cross-agent spawning from the orchestrator?
I am designing a multi-agent architecture in OpenClaw with multiple Telegram bots and need clarification on the correct configuration approach.
Current Setup
I have created four Telegram bots:
Larry
Bob
Sentinel
Alfred
My goal is the following:
Each bot should correspond to a dedicated agent.
Each agent should use a different model configuration.
Larry should act as the orchestrator.
Larry must be able to spawn and coordinate the other agents.
At the same time, I want to be able to directly interact with each bot independently in its own Telegram chat.
Desired Behavior
When I write to Larry's Telegram bot, Larry responds and can delegate tasks to Bob, Sentinel, or Alfred.
When I open Bob’s Telegram chat, I want to talk directly to Bob (without going through Larry).
Same for Sentinel and Alfred.
Each agent must remain isolated in its own Telegram context.
No message leakage between agents.
Model configurations must remain independent per agent.
Questions
What is the correct configuration pattern to:
Bind one Telegram bot per agent?
Ensure strict session isolation per bot?
Allow orchestration while still permitting direct conversations?
Should each agent define:
Its own Telegram transport?
Its own workspace?
Its own binding rules?
What is the recommended architecture for:
Orchestrator agent with delegation privileges
Specialist agents accessible both directly and via orchestration?
Is there a minimal working example of a configuration that supports:
Multiple bots
Multiple agents
Independent models
Cross-agent spawning from the orchestrator?
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
131,281Members
Resources
Similar Threads
Was this page helpful?
