What it does
5 features
Agent Zero · Self-hosted work agent
Product and access
The A0 Launcher simplifies Agent Zero setup and is an entry point to the same product.
Core features
5 features from provider sources · reviewed on 12 October 2026. Availability and limits are listed below.
Dedicated Linux environment
Agents work with terminals, files and desktop tools inside a Docker environment.
Source 2
Usable AI models
Open choice
You set a main and a utility model yourself: with your own key, OpenRouter, an account sign-in such as ChatGPT, or an OpenAI-compatible endpoint. Models run locally through Ollama, LM Studio, llama.cpp, oMLX or vLLM. Plugins add further providers.
Model families named
- OpenAI (GPT)
- Anthropic (Claude)
- Google (Gemini)
- xAI (Grok)
- DeepSeek
- Mistral
- Moonshot (Kimi)
- Z.ai (GLM)
- Other models can be connected freely
- Model choice
- You choose
- Your own API key
- Yes
- Router or custom endpoint
- Yes
- Local models
- Yes
The rating covers the generally available main access. Documented · Sources accessed on
Agent Zero: First-Run Onboarding · Agent Zero: Installation · Agent Zero: model_providers.yaml
Does it fit my task?
2 practical tasks · 3 requirements and limits
A good fit if …
- you can run Docker on your own machine or server and manage the instance yourself
- you want the agent to work in a Linux workspace with a browser and files
- you want to schedule tasks via cron or at fixed times
Less suitable if …
- you want a hosted product without installation and updates
- you need management with roles and administrator approvals; the reviewed pages describe none
Editorial assessment based on provider documentation. Not a product test.
Practical tasks
Editorial examples based on documented features. No product tests performed.
Process documents
- Your input
- Shared files and a clear task.
- Possible result
- A structured overview of the supplied documents.
- What to check
- Check statements, files and granted access.
Delegate research
- Your input
- A topic and a bounded set of sources.
- Possible result
- A research draft with traceable sources.
- What to check
- Open sources and review conclusions.
A structured overview of the supplied documents.
Your possible prompt
Compare three providers for [task]. Use their official documentation. Show differences, sources, and open questions in a table.
Adjust the task, data and permissions.
Requirements and limits
Not yet confirmed
Reliability in your workflow
Information is based on provider sources; success rates for specific workflows have not been measured here.
Source 2
Can I use it?
With requirements
Europe & regional availability
Installation instructions and model prerequisites reviewed. The assessment covers the local operating path described; complete country and feature availability of all connected services remains open.
Reviewed scope: Local software and documented self-hosting; external services assessed separately
Reviewed on:
With requirements
Self-hosting with own model access
Agent Zero documents operation in your own Docker container on a computer or server. A model provider is configured after startup; its regional availability must be checked separately.
Region, reason and source
Region covered: Europe: own computer or server; model services are country-specific
Documented reason: Technical assessment based on installation and provider selection. These sources do not establish a legal cause for regional restrictions.
Accounts & permissions
Based on provider documentation. Permissions depend on the account, integration and organization policies. Review the permissions actually requested before use.
Configuration covered: Agent Zero in a Docker container on your own machine or server with its web interface
Sources reviewed:
Show all 3 points
Documented with conditions
Data access & team sharing
The direct Docker command binds the instance to your computer only; the page does not say this for Launcher or other routes. With host access, file edits and commands affect real files on your computer. Projects bundle instructions, files, memory, variables and secrets.
Evidence · Evidence
Memory and conversation continuity
3 of 6 criteria documented
The memory dashboard supports search and maintenance; provenance distinguishes memories from knowledge imports.
Native memory and separate imported knowledge
Sources accessed:
Documentation by criterion
StoragePartly documented
Storage and scope
Memory resides in selectable directories and areas; imported knowledge fragments have separate provenance. The specific backend type is not named here.
SelectionPartly documented
Selection and timing
Persistent agent memory is documented; exact automatic extraction triggers are not specified in the management guide.
Edit & deleteDocumented
Inspect, edit and delete
The dashboard shows text and metadata, supports editing, deletion and export, and allows selecting the memory directory.
Access boundariesDocumented
Agent and channel boundaries
Global and project memory directories and areas such as fragments and solutions can be distinguished.
RetrievalDocumented
Retrieval
Memory search offers a similarity threshold, result limit and area filters; imported knowledge can be assessed separately.
New conversationsPartly documented
Session reset and compaction
Memories remain available across chats. The exact interaction between conversation compaction and memory is not described here.
Cost and effort
1 plan · Prices as of 12 October 2026
Prices and included usage
Pricing checked
MIT-licensed; model and hosting costs are additional.
| Plan | Price | Included |
|---|---|---|
| Open Source | USD 0 license cost | MIT license; hosting and model usage are separate. |
- Cloud models may incur API costs; local models require your own compute.
Setup
Install Docker, start an instance and configure a cloud or local model.
Operation & oversight
Based on public provider documentation. Assessments are editorial inferences from the stated properties. Actual time requirements and reliability have not been measured.
Deployment covered: Agent Zero as a long-running Docker instance with its task scheduler
Sources reviewed:
Editorial assessment
You run Docker, updates and backups yourself and protect the interface before it is reachable beyond your machine. Failed tasks you have to check actively in the scheduler.
Show all 6 points
Documented
Files and working state
Persistent storage keeps settings, chats, memory, projects and plugins when a container is replaced. If you choose no persistent workspace, deleting the container can also delete its data.
EvidenceOpen in reviewed sources
Sign-ins
The reviewed pages do not describe how long the web sign-in remains valid.
EvidenceDocumented
Starting further runs
The task scheduler supports recurring tasks via cron, planned one-off times and manual-only tasks. A run uses its own chat context (the default for tasks created in the UI) or a shared one.
EvidenceDocumented with conditions
Detecting errors
A failed task shows the Error state, and Last Result shows output or error from the last run. As best practice the page advises alerting on failures and mentions execution logs without saying where they are stored.
Evidence
Evidence and open questions
4 sources · 1 open point
Open points
The reviewed sources contain no reliable information on these points. A gap says nothing about the product.
Requirements and limits: not yet confirmed
Sources and evidence
Official product documentation reviewed. Workflows and assessments are editorial; hands-on product tests are outside the scope of this directory.
Editorial team: Michael Meierhoff · sources reviewed on . Provider information is not independent evidence of performance.
Provider
Provider of the open Agent Zero framework.
- Operator
- Agent Zero, s.r.o.
- Organization
- Company
- Website
- www.agent-zero.ai
Changes to this profile 1
New in the catalog: a self-hosted work agent with a Linux environment, browser, files and subagents.
These notes describe editorial changes. Source and pricing review dates appear in their respective sections.