Practical overview
When it fits
Use apowerb when you need to build and operate agents that connect to business tools or internal data, especially when runs should be scheduled or triggered by incoming email.
Who it is for
- Developers building agent applications and workflows.
- Teams connecting AI agents to business systems and internal data.
- Organizations seeking to self-host agent infrastructure with Docker Compose or Helm.
Documented uses
- Compose agents into sequential, parallel or loop workflows, including sub-agents.
- Ground agent answers in files, URLs, databases or other connected data using RAG.
- Query databases in natural language through Text-to-SQL.
- Trigger workflows from schedules or new Gmail and Outlook messages.
- Connect agents to tools such as Google Workspace, Microsoft 365, CRM, ERP and SQL systems.
Requirements & limits
The repository describes the core as open-source under Apache-2.0 and documents self-hosting with Docker Compose or Helm. Users bring their own model key. Several enterprise modules are sold separately.
- Agents are defined through a web interface or REST API, rather than a canvas workflow builder.
- The framework is built on FastAPI, Google ADK and LiteLLM; model access depends on a provider supported through LiteLLM and a user-supplied key.
- RAG, Text-to-SQL, business integrations and unattended triggers are relevant when workflows need data access or action beyond chat.
- Billing, identity-provider sign-in, MFA, agent evaluation and the supervision screen are identified as separately sold enterprise modules.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Start apowerb with Docker Compose
Set up the included Docker Compose stack and open the local interface.
Before you start
- Git and Docker Compose are available.
- A model API key is needed for agents to answer; it can be added later in the interface or declared in .env.
Steps
Clone the apowerb-hosting repository and change into its directory: the command below.
git clone https://github.com/apowerb/apowerb-hosting.git && cd apowerb-hostingCopy the example environment file and generate stack secrets:
cp .env.example .env && ./scripts/generate-secrets.sh.Start the stack: the command below.
docker compose -f docker-compose/docker-compose.yml --env-file .env up -dOpen the interface at http://localhost:3000; the API is on port 8000. Add your own model key in the interface or configure
DEFAULT_LLM_MODELandDEFAULT_LLM_API_KEYin.envto make a model available.
Expected outcome
The Compose stack runs with PostgreSQL included, and its interface is available at localhost:3000. Without a configured model, the model does not appear in the list and agents cannot answer.
Guide limitations
- This documented quick start uses Docker Compose and a local URL; it does not document deploying the stack to a public environment.
- The framework cannot provide a model API key; bring your own key to use a model.
- The repository README points to the separate apowerb-hosting repository for this stack.
Install apowerb from source
Install the Python project into a virtual environment and prepare its environment file.
Before you start
- Python 3.10 or later.
- UV, installed with pip.
- A PostgreSQL database you provide; source installation requires PostgreSQL configuration.
Steps
Install UV and create a virtual environment: the command below followed by the command below.
pip install uvuv venvActivate the environment: on Windows, run
.venv\Scripts\activate; on macOS or Linux, runsource .venv/bin/activate.Install project dependencies: the command below and then the command below.
uv syncuv pip install .Create the environment file:
cp .env.example .env. For this own-PostgreSQL installation, configure the requiredDB_HOST,DB_NAME,DB_USER,DB_PASSWORD, andENCRYPT_KEYsettings in the environment.
Expected outcome
The project dependencies are installed in the activated virtual environment, and an environment file is prepared with the settings needed for a source installation.
Guide limitations
- This installation path expects an external PostgreSQL database; the Compose quick-start stack instead includes PostgreSQL.
- The instructions here do not document starting or running the application after installation.
Compare options for Agent orchestration
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| Apowerb | Compose agents into sequential, parallel or loop workflows, including sub-agents. | CLI | The repository describes the core as open-source under Apache-2.0 and documents self-hosting with Docker Compose or Helm. Users bring their own model key. Several enterprise modules are sold separately. |
| Agenite | Define typed tools with JSON Schema validation and error handling. | CLI | |
| Veldra | Create an agent with a policy, tools, skills, knowledge base and reasoning method. | Self-hosted | The documented deployment is self-hosted and uses a sign-in gate after a first-run administrator account is created. |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 7 October 2026
- Original website title
- apowerb/apowerb · Official homepage title · retained in its source language