Module 1
Understanding INEMACCBOT
What the system is and why it is not just a Telegram bot. Presentation of the six components and the central rule that guides the entire course.
Next cohort · coming soon
Practical course
Promoavatar as a sample project
Course INEMACCBOT in practice: building a campaign with Promoavatar
Learn the system’s mental model and then follow a real campaign being produced end‑to‑end — from topic to reel delivery.
The mental model
INEMACCBOT is not just a Telegram bot: it is the gateway and orchestrator that receives commands, creates jobs, controls queues and maintains flow state. Everything the course teaches is built on six components and each one’s role.
Bot→Flow→Queue→Worker→Agent→Skill
The central rule: skill executes · flow coordinates · queue organizes · worker processes · agent operates the AI · bot connects the user to the system.
The six components
Bot
The entry point
Connects the user to the system: receives the command on Telegram and returns the result.
Flow
Who coordinates
Monitors the process state and decides which job should be created next. It does not execute: it coordinates.
Queue
Who organizes
Keeps the jobs waiting for execution, by class, and maintains order under load.
Worker
Who processes
Pulls the job from the queue and actually executes it, reporting success or failure back to the flow.
Agent
Who operates the AI
Runs the model: interprets the request, chooses the path and produces the requested content.
Skill
Who executes
The concrete capacity at the edge — transcription, dubbing, image, explanatory video, reel.
The two execution modes
Simple task — direct path
Bot → Queue → Worker → Agent → Skill → Result
A task that starts and ends in a single pass does not need memory: it enters the queue, is processed and returns the result.
Process with steps — goes through a flow
Bot → Flow → (Queue → Worker → Agent → Skill)ⁿ → Result
When the process needs to remember where it stopped — pause, wait for human approval, resume — it requires a flow with a state machine.
The program — 10 modules
Module 2
Skill or Flow?
The two possible paths. A simple task goes straight through Bot → Fila → Worker → Agente → Skill; a process that needs to remember where it stopped must use a flow. Examples of already documented skills: transcription, dubbing, image, explanatory video and reel.
Module 3
Promoavatar as a real flow
Here comes the practical case. The student executes /promoavatar <assunto>: the system creates specific texts for each audience, pauses processing, waits for the avatar step in HeyGen, receives approval and then continues reel production.
Module 4
Inside a campaign
The didactic flow of the complete campaign, step by step. The important point is to see that the flow does not execute everything alone: it monitors the state and decides which job to create next, while the queues organize the jobs waiting for execution.
Module 5
Audiences and personalization
How a single campaign produces different content for several audiences. In Promoavatar, each audience receives its own script, and the destination channel can be set by the flow.json.
Module 6
Human intervention is part of the flow
A central concept: automation does not mean eliminating the human. In Promoavatar there is a deliberate pause to review texts, produce and record avatars and approve the batch before the following phases.
Module 7
State, resumption, and errors
Why a complex process needs operational memory. The flow knows what finished, what is executing, what failed and where it must continue after an interruption.
Module 8
Operating the system via Telegram
Practical part with everyday commands: /skills, /fluxos, /fila, /status, /jobs, /cancelar, /refazer, /furar, /pronto and /limpar.
Module 9
How to create new flows
After understanding Promoavatar, it becomes clear that the same mechanism serves PromoCurso, PromoEvento, PromoProduto, course production and other campaigns — all sharing the same queue classes.
Module 10
Final project
The student creates a complete campaign, from theme to distribution. Upon completion, they will not only have learned commands: they will have understood how to design automations on INEMACCBOT.
A campaign, end to end
This is the path Promoavatar follows — and you will follow it inside, seeing the state change at each step:
subject→scripts→approval→HeyGen avatar→locate video→reel→delivery
In amber, the steps where the human enters the flow — the pause is designed, not a defect.
The commands you will use
View and follow
| Command | Purpose |
|---|---|
/skills | List available capabilities |
/fluxos | View flows and their states |
/fila | Inspect what waits for execution |
/status | Overall system status |
/jobs | Ongoing and completed jobs |
Intervene
| Command | Purpose |
|---|---|
/cancelar | Interrupt a job or campaign |
/refazer | Reprocess a failed step |
/furar | Prioritize a job in the queue |
/pronto | Approve the batch and release the resume |
/limpar | Clear queues and residual state |
What you take away at the end
- The complete mental model of INEMACCBOT
- A Promoavatar campaign executed by you
- Clarity on when to use skill and when to use flow
- Mastery of operation commands in Telegram
- Understanding of queue, worker, agent and state machine
- Ability to design new flows on the same mechanism
Skill executes, flow coordinates, queue organizes, worker processes, agent operates the AI and bot connects the user. Once understood, you are no longer using the system: you are designing on top of it.
Next cohort soon
Join the INEMA Community to learn about the opening — or talk directly to @apoioinema on Telegram.