Built for real workflows

AI bots with channels, memory, tools, and human backup.

Each Botmain bot combines a task-specific AI agent, channel integration, controlled knowledge, usage metering, and setup support - so businesses can automate work without building their own AI infrastructure.

Section 01

From message to metered response

Every bot message passes through a controlled workflow: channel input, agent instructions, model routing, approved knowledge, response delivery, and credit metering.

Step 01 · Active

Channel

User message

"Can you send an invoice reminder?"

Bot response

Reminder sent30 credits

Section 02

Every bot is configured for a specific job

Each Botmain persona is not a generic chatbot - it has a defined role, tools, workflows, and fallback rules.

Role & persona

A defined identity with a clear job to do.

Task-specific instructions

Prompts tuned to the workflows the bot supports.

Allowed actions

Only the tools each bot needs - nothing more.

Channel configuration

Wired to the messaging platforms the business uses.

Fallback rules

Escalation paths for anything out of scope.

Credit usage rules

Metering profile matched to the workload.

Business knowledge

Optional grounding in approved documents.

Human handoff

Clear paths from bot to a real person.

Examples in the current lineup

Bookkeeper

Expense logging, invoice reminders, cash-flow summaries.

Shop assistant

Product questions, availability, guided recommendations.

Travel planner

Trip planning, bookings, and customer requests.

Receptionist

Common questions and routing to the right person.

Social media

Post drafts, captions, and content ideas.

Section 03

Model layer and routing

Botmain can route tasks through different AI models depending on the workload. Simple replies can use faster, lower-cost models, while longer reasoning, research, or document-heavy tasks can use more advanced models.

Quick replies
Fast response model
Low credit usage
Standard tasks
General-purpose AI model
Standard credit usage
Long context / deep task
Higher-capability reasoning model
Higher credit usage
Document generation
Model + structured output
Higher credit usage
Multi-step research
Agentic workflow + reasoning model
Highest credit usage
Premium reasoning
Advanced model route
×2 multiplier

Exact model providers and configurations may vary by bot, use case, and availability.

Section 04

Bots answer from controlled knowledge

Bots can be grounded in website content, FAQs, business documents, product information, service descriptions, or approved internal knowledge - to keep answers aligned with the business and reduce hallucinations.

Website content

FAQs

Product catalogue

Business rules

Setup instructions

Human support fallback

Guardrail: Bots should not guess when the answer is not available. If a question is outside the approved knowledge or too complex, the bot can route the user to human support.

Section 05

Works where customers already message

Each channel uses a dedicated connector that receives messages, hands them to the right bot, and sends the response back on the same thread.

WhatsApp

Supported

Business API via approved messaging provider.

Telegram

Supported

Bot token connected to your Telegram workspace.

Website chat

Planned

Embed a chat surface directly on your site.

Future channels

Roadmap

Additional connectors as demand grows.

Setup may require connecting the business account, bot token, or messaging provider. Our team handles it.

Section 06

Bots can do more than reply

Depending on the use case, a bot may connect to tools or systems to take action - not just chat.

Create reminders
Summarize conversations
Generate documents
Log expenses
Check product information
Prepare reports
Hand off to a human
Trigger internal workflows

Guardrail: Tool access is configured per bot and only enabled where appropriate.

Section 07

How usage becomes BotCredits

Every bot action consumes credits. Simple replies cost fewer credits; long, complex, or multi-step tasks cost more. 1 BotCredit = $0.01.

Quick reply8 credits
Standard task30 credits
Deep task150 credits
Document generation400 credits
Multi-step research600 credits
Premium reasoning×2 multiplier

Section 08

Designed with control and security in mind

Security controls are designed around least-privilege access, controlled integrations, and clear escalation paths.

Least-privilege access

Bots and integrations only get the permissions they need to run.

Human review for sensitive actions

Actions that touch billing, customers, or private data can require confirmation.

Purpose-bound data use

Customer data is used to operate the bot and generate responses.

Operational logging

Usage data may be used to monitor performance and improve reliability.

Checkout provider handles payments

Raw card details are processed by the checkout provider, not stored by Botmain.

Controlled integrations

Every channel and tool is wired in through a reviewed connector.

Section 09

Human support when automation is not enough

Bots should not handle every situation alone. If a request is too complex, unclear, unsupported, or sensitive, the user is routed to a human.

Talk to a human

We help set up bots, connect channels, and handle edge cases.

Contact support

Section 10

Technical FAQ

Straight answers for buyers evaluating the stack.

Ready to put a bot to work?

Meet the team, pick a role, and our engineers handle setup end-to-end.