Practical overview
When it fits
Use FleetQ for multi-agent processes that need branching, loops, specialist roles, external tools, scheduled or event-driven execution, human approvals or operational records.
Who it is for
- Developers building multi-agent applications
- Platform and AI-operations teams managing agent workflows
- Teams requiring human review, model-provider choice and execution records
- Organisations evaluating self-hosted AGPL-licensed orchestration software
Documented uses
- Design reusable visual workflows for multi-step agent processes
- Coordinate specialist agents in structured crews
- Connect agents to external capabilities through MCP
- Route selected actions through human approval and risk policies
- Monitor executions, budgets and audit trails
Requirements & limits
The project is presented as open-source under AGPL-3.0 and supports self-hosted deployment through Docker Compose. The repository also references FleetQ Cloud, but the supplied material does not establish its current pricing or availability.
- Self-hosting is central to the project’s documented deployment model.
- The visual DAG builder supports agent, conditional, human-task, switch, dynamic-fork, do-while, compensation and sub-workflow nodes.
- Provider choice spans Anthropic, OpenAI, Google and several local or OpenAI-compatible systems.
- MCP integration is a core operating model, with the repository describing more than 670 tools across over 60 domains.
- Human approval and risk-policy controls matter when automated actions require review.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Set up FleetQ for self-hosted agent orchestration
Prepare the documented self-hosted FleetQ platform and its supported model integrations.
Before you start
- A server environment capable of running Docker Compose.
- API credentials or access for a supported provider or local model system, such as Anthropic, OpenAI, Google, Ollama, LM Studio, vLLM, Codex or Claude Code.
- Familiarity with the AGPL-3.0 licensed project and its repository.
Steps
Obtain the FleetQ source from the public GitHub repository.
Use the project’s Docker Compose deployment approach for the self-hosted installation.
Configure the chosen supported model provider or local model integration.
Create agents and assemble them into visual DAG workflows with any required human-task nodes.
Use the web interface to monitor workflow activity and review approval or audit information.
Expected outcome
A self-hosted FleetQ environment configured for building and operating multi-agent workflows.
Guide limitations
- The supplied material documents the deployment approach and supported integrations but does not provide the complete command sequence or verify a successful installation.
- Model access, credentials and external service configuration remain dependent on the selected provider or local system.
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 |
|---|---|---|---|
| FleetQ | Design reusable visual workflows for multi-step agent processes | Self-hosted, Server, Web | The project is presented as open-source under AGPL-3.0 and supports self-hosted deployment through Docker Compose. The repository also references FleetQ Cloud, but the supplied material does not establish its current pricing or availability. |
| 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
- 4 October 2026
- Original website title
- escapeboy/agent-fleet-o · Official homepage title · retained in its source language

