DMJBot

Your AI Employee That Never Sleeps.
DMJBot works while you live. Automate emails, workflows, and routine tasks with an AI assistant that runs 24/7

[DMJBot @domyjobbot]

DMJBot initiativity lets your AI assistant proactively review context, prepare updates, and start helpful work within limits you control.

https://t.co/R9HPVG1mMp

[DMJBot @domyjobbot]

The initiative of AI assistant is the unique feature in DMJBot. We are ready to release it. Soon we will share more details!

[DMJBot @domyjobbot]

Delegation without visibility is a black box.

In DMJBot every delegated run shows up as a subsession: collapsed it is one line, expanded it is the briefing, every tool call, and the result.

https://t.co/8YwWnG1YaH

#AIAgents #DMJBot

[DMJBot @domyjobbot]

The work often has to happen where the files and credentials already are, not in the cloud.

So DMJBot defines an agent as an MCP server exposing `run_agent` and `cancel_task`. That is the whole contract — Claude Code, Copilot CLI, aider, Codex, a shell script.

#AIAgents #MCP

[DMJBot @domyjobbot]

A plain agent loop has one serious weakness: every round costs more than the last, because every tool result makes the context bigger.

Delegation is the fix. The subagent does the messy twelve-call part on the side and returns one line.

[DMJBot @domyjobbot]

The obvious problem with a local model: your machine sleeps, and the chat stops.

So DMJBot falls back. Primary, fallback, secondary, secondary fallback — when your Ollama box is unreachable, the next model answers.

https://t.co/dWobKMqoUj

#LocalLLM #Ollama #AIAgents

[DMJBot @domyjobbot]

Your laptop has the model. Your cloud assistant has the uptime. Your phone is somewhere else entirely.

DMJBot now connects the three: run open-weight models with Ollama at home, use them from your cloud instance, reach it from the phone.

[DMJBot @domyjobbot]

When you type, you compress. You drop the background that felt obvious to you — which is usually the exact context the assistant needed.

When you talk, you don't compress. The model handles the rambling fine.

https://t.co/FM7z49pJ0q

#AI #AIAgents

[DMJBot @domyjobbot]

An idea arrives while you are away from the laptop.

You are not going to type three paragraphs standing in a queue. So you type one line, and your assistant gets one line of context.

Talk instead, and it gets all of it.

[DMJBot @domyjobbot]

A PR opens. CI fails. A customer asks for an update.

Three events, or one situation: the release may be blocked?

Kateryna on perception in AI harnesses:
https://t.co/BectNhH9OX

[DMJBot @domyjobbot]

AI systems should not only process events.

They need perception: connecting signals and understanding what is happening in context.

Kateryna's post:
https://t.co/BectNhH9OX

[DMJBot @domyjobbot]

@Joi2James Freedom

https://t.co/fjGBQDFiEJ

[DMJBot @domyjobbot]

We are integrating https://t.co/N0IPpzDoyI with https://t.co/fjGBQDFiEJ . We already can see a lot of opportunities. From one side a dmjbot instance can be a manager of external tasks, from other you can pull BUZZ chats in a context of your assistant projects

[DMJBot @domyjobbot]

@trikcode What about comparing Kimi to Deepseek? did anybody tested this?

[DMJBot @domyjobbot]

@trikcode Which harness do you use for coding with this model?

[DMJBot @domyjobbot]

One big benefit of external memory: centralization.

For multi-machine or multi-assistant workflows, keeping memory in one place can simplify operations.

That is why we support @mem0ai in DMJBot.

#AIAutomation #Mem0 #DevOps

[DMJBot @domyjobbot]

If your assistant memory strategy needs to scale, external memory is worth a look.

Thanks @mem0ai for building this.

#AIAgents #Memory #Mem0 #DMJBot

[DMJBot @domyjobbot]

Memory for AI should be a choice.

With DMJBot you can run:
- built-in internal memory
- external memory with @mem0ai

Different teams, different needs.

#AI #Mem0 #ProductEngineering

[DMJBot @domyjobbot]

Big win: our AI assistant brain can run anywhere with @Docker.

Laptop, cloud VM, or on-prem server, same containerized stack.

https://t.co/cdWQpjgXes

#Cloud #AIAgents #Docker

[DMJBot @domyjobbot]

We support @mem0ai in DMJBot as an external memory option.

Users can keep assistant memories centralized instead of only using local internal memory.

#AIAgents #Memory #Mem0 #DMJBot