# Moltbot Beginners Setup Guide and Alternative for Free?

Source: https://www.youtube.com/watch?v=PleO6KKkfVo
Recap page: https://rapidrecap.app/video/PleO6KKkfVo
Generated: 2026-01-29T20:36:38.373+00:00

---
## Quick Overview

The video provides a comprehensive tutorial on setting up and running a custom AI agent using PraisonAI via Slack integration, utilizing Ollama locally for model inference, covering Slack API configuration, agent definition in YAML, environment setup, and execution via a single command.

**Key Points:**
- The setup involves installing the PraisonAI bot via a single command: `curl -fsSL https://molt.bot/install.sh | bash` (02:00).
- Slack configuration requires creating an app on the Slack API Console, configuring OAuth & Permissions with specific scopes (e.g., chat:write, app_mentions:read, im:history), enabling Socket Mode, and subscribing to bot events (03:12 - 05:00).
- The agent's behavior is defined in an `agents.yaml` file, specifying roles, goals, tools (like `search_web`), and tasks for each agent (02:48).
- The bot is run using a command that specifies the channel type (slack), bot/app tokens, auto-approval flag, the local Ollama model (e.g., `--model ollama/ministral-3`), and the agent definition file (e.g., `--agent agents.yaml`) (08:34).
- The initial setup includes installing missing skill dependencies via an interactive prompt, such as `bear-notes` (06:08).
- The system successfully demonstrates agent interaction in a Slack thread, with the agent responding based on the defined local Ollama model (07:30).
- The tutorial highlights that agents can be run locally using Ollama for free, providing full control over the process (08:25).

![Screenshot at 00:05: The user issues a command \(@Praison AI list files in ~/moltbot\) in Slack, prompting the custom agent to execute a file listing command on the local system, demonstrating tool execution capability.](https://ss.rapidrecap.app/screens/PleO6KKkfVo/00-00-05.jpg)

**Context:** The video demonstrates how to deploy and run custom AI agents built with the PraisonAI framework, integrating them with the Slack messaging platform. This setup bypasses cloud services by running the LLM inference locally using Ollama, which requires setting up tokens, defining agent capabilities via YAML, and configuring Slack permissions for bot interaction.

## Detailed Analysis

The tutorial guides the user through setting up a custom AI agent using PraisonAI integrated with Slack, leveraging Ollama for local, free LLM execution. The process begins with installing the Moltbot installer via a one-liner curl command (02:00). Next, the user configures the Slack application by navigating the Slack API console (03:41). This involves creating the app, configuring OAuth & Permissions by adding necessary Bot Token Scopes (`chat:write`, `app_mentions:read`, `im:history`), enabling Socket Mode (04:24), and subscribing to bot events like `app_mention` and `message.im` (04:56). Concurrently, the agent's structure is defined in an `agents.yaml` file, detailing roles like 'Researcher' with the `search_web` tool, and a 'Summarizer' agent (02:48). The environment variables for Slack tokens must be exported in the terminal (08:15). Finally, the bot is launched using a comprehensive CLI command specifying Slack, the exported tokens, the local Ollama model (`--model ollama/ministral-3`), and the agent configuration file (`--agent agents.yaml`) (08:56). The successful execution is verified by sending a message in Slack, receiving a response generated by the locally running model (07:30). The demonstration concludes by showing the agent successfully executing a local command (`list files in ~/moltbot`) (09:49), confirming the full local integration.

### Moltbot Installation & Setup

- Installation performed via one-liner: `curl -fsSL https://molt.bot/install.sh
- bash` (02:00)
- Gateway service installation confirmed with Node runtime (02:12)
- Hook configuration skipped for initial setup (06:27).

### Slack App Configuration

- Create Slack App from scratch via Slack API Console (03:15)
- Configure OAuth & Permissions: Add scopes including `chat:write`, `app_mentions:read`, and `im:history` (03:51)
- Enable Socket Mode and generate an App Level Token with `connections:write` scope (04:30).

### Agent Definition (agents.yaml)

- Defines multiple agents, such as 'searcher' using `search_web` tool and 'summarizer' for content condensation (02:48)
- Agents are configured using the framework 'praisonai' (10:00).

### Running the Bot

- Bot started using `praisonai bot slack` command, passing both Slack tokens and specifying the local Ollama model (`--model ollama/ministral-3`) and agent file (`--agent agents.yaml`) (08:56).

### Local LLM Integration

- Ollama downloaded and run locally using `ollama run ministral-3` (08:44)
- The agent successfully responds via Slack using the locally hosted model (07:30).

### Agent Capabilities Demonstration

- Agent confirms it is a program without feelings but is ready to help (07:29)
- Agent successfully executes a tool command (`list files in ~/moltbot`) via Slack (09:49).

![Screenshot at 00:00: The initial Slack interface showing a summary of previous questions asked to the Praison AI bot \(00:04\).](https://ss.rapidrecap.app/screens/PleO6KKkfVo/00-00-00.jpg)
![Screenshot at 02:00: The Moltbot installer one-liner command being executed in the terminal, which installs necessary dependencies \(02:12\).](https://ss.rapidrecap.app/screens/PleO6KKkfVo/00-02-00.jpg)
![Screenshot at 03:12: The step-by-step guide for Slack Platform Setup, detailing the creation of the Slack App and OAuth configuration \(03:22\).](https://ss.rapidrecap.app/screens/PleO6KKkfVo/00-03-12.jpg)
![Screenshot at 08:15: The terminal showing environment variables being exported for Slack tokens \(SLACK\_APP\_TOKEN and SLACK\_BOT\_TOKEN\) before running the bot \(08:22\).](https://ss.rapidrecap.app/screens/PleO6KKkfVo/00-08-15.jpg)
![Screenshot at 07:41: The thread view in Slack showing the AI agent responding to a query about its built-in skills, listing tools like weather, slack, github, and notion \(07:47\).](https://ss.rapidrecap.app/screens/PleO6KKkfVo/00-07-41.jpg)