AI tasks for Digital.ai Release

Create AI-powered automation workflows in Release!

This integration provides tasks for Release to interact with LLMs like OpenAI and Gemini and interact with any MCP servers.

• Create AI agents that perform complex tasks using MCP tools
• Call services directly using MCP, without writing any code
• Embed AI prompts in your Release workflows
• Start interactive chats with the LLM of your choice
• Connect to Gemini and OpenAI-compatible providers
• Connect to any MCP server
• Mix and match MCP servers and reasoning models per task, to get the right tool for the job

Tasks

AI: Agent

Creates an AI agent that can use MCP tools to accomplish tasks.

AI: Prompt

Connect to an AI server and invoke a single prompt. Supports Gemini and OpenAI-compatible LLM providers.

AI: Chat

Interactive chat with an LLM, inside the task's Activity section.

MCP: Call tool

Connect to any MCP server and invoke its tools. No LLM reasoning, just direct tool calls.

MCP: List Tools

Lists available tools on an MCP server

Demo templates

This repo comes with example templates showing how to use the tasks.

1. MCP Examples -- Demonstrates the MCP tool tasks that allows you to interact with 3rd party servers without having to write an integration plugin. (no LLM involved).
2. Prompt examples -- Demonstrates the AI Prompt task with different model providers (Gemini, OpenAI, etc).
3. MCP + Prompt examples -- Demonstrates how to combine MCP and Prompt tasks to build AI-powered workflows.
4. Agent examples -- Demonstrates the AI Agent task that combines LLM prompting with multiple MCP servers.
5. Interactive Chat example -- Demonstrates the interactive AI Chat task.

See the Build & Run section on how to install the examples.

---

Build & Run

This section explains how to set up a local Release instance with the plugin and upload the example templates.

Prerequisites

You need to have the following installed in order to develop Python-based container tasks for Release using this project:

• Python 3
• Docker

Configure your hosts file

Add the following to /etc/hosts or C:\Windows\System32\drivers\etc\hosts (sudo / administrator permissions required):

127.0.0.1 host.docker.internal
127.0.0.1 container-registry

Start the development environment

This project comes with a Docker environment that sets up a local Release instance with all the required services.

Running the environment with this plugin is a three-step process:

1. Start the Release Docker environment
2. Build and publish the plugin
3. Upload the demo templates

When developing the plugin, typically you would just do step 2 after making code changes. The new version of the plugin will be picked up without having to restart the server.

1. Start the Release environment:

The Release environment is defined in dev-environment/docker-compose.yaml.

Stopping

Have Docker running and launch the environment with:

docker compose -f dev-environment/docker-compose.yaml up -d --build 

It takes a while to start up. You can see that the Release server is running when the digitalai-release-setup container has terminated.

Check if you can log in with admin/admin at http://localhost:5516.

Stopping

After the demo, you can stop the environment with:

docker compose -f dev-environment/docker-compose.yaml down

2. Build & publish the plugin

The build.sh script will build the plugin container, publish it to the local registry and install it to the local Release instance.

Run the build script

Unix / macOS

sh build.sh --upload

Windows

build.bat --upload

3. Install example templates

Set up credentials

The sample templates come with examples to connect to various MCP servers and LLM providers.

Put your API keys in setup/secrets.xlvals.

Use the example file as a base:

cp setup/secrets.xlvals.example setup/secrets.xlvals

Then edit the file and add your keys.

Upload templates

Run the following command to upload the demo templates to the local Release instance:

./xlw apply -f setup/mcp-demo.yaml

The templates will be uploaded to a new AI Demo folder.

---

Demo time!

1. Log in to http://localhost:5516 with admin/admin
2. Go to the AI Demo folder
3. Go the Templates section and run the examples

👉Add your favorite MCP Server or LLM provider under Connections and build your own example!

---

Developing

Run the tests

The tests are integration tests and need API keys to run. Put your API keys in .env file in the root of the project:

GEMINI_API_KEY=<key>
OPENAI_API_KEY=<key>
DAI_LLM_API_KEY=<key>

Run the tests with the command