6 capabilities
Knowledge and memory
Answers from your documents with citations, and what the agent should remember.
- Knowledge search
- Skills
- Context
- Memory
- Memory (mem0)
- Past conversations
Open sourceSelf-hosted · Apache-2.0 · built on Pydantic AIPut AI to work
Put AI to work
across your company.
The open-source agent layer for shared agents, company knowledge and automation, on infrastructure you control.
Free to run yourself under Apache-2.0. Rolled out by Vstorm when you want help.
Works with the models you already use:
AgenticOS gives every team AI teammates: agents that work with files, run code, browse the web and use your company’s tools and knowledge. Built in a browser, shared with colleagues, run on servers you own.
- built-in capabilities, switched on per agent
- 26
- model providers, hosted or local
- 27
- places one agent answers
- 8
- MCP servers in the gallery
- 4,700+
AI teammates
Agents with a name, a job and a place on the team.
Each one has its own instructions, tools, knowledge and budget, and colleagues reach it in chat, in Slack or from the API. These six work in the demo organization.
Chief of Staff
Hands parts of a task to colleagues and writes up the brief.
- Subagents
- Planning
Your experts, multiplied
Your best people’s judgment, in agents you own.
Growing a company used to mean hiring more experienced people. Here their procedures become skills, their standing facts become context, and every agent works from both. Change a rule once and every agent that reads it answers with the new one.
SHow it is doneskills/invoice-dispute.md
- 1. Match the invoice to the PO and the delivery note.
- 2. List every line that differs, with the amount.
- 3. Draft the reply to the supplier in our tone.
CWhat is true todaycontext/procurement-thresholds.md
- Disputes up to €10,000: procurement decides.
- Above €10,000: the CFO signs off.
- Preferred suppliers: see the 2026 list.
A dispute above €10,000 waits for the CFO
Examples. Your procedures and figures go in their place.
How it works
From a brief to the boardroom, in seven steps.
One agent, from its first sentence to a dashboard the board opens: built, taught, connected to your tools, asked in the chat and in Slack, put on a schedule and shared.
Brief it like a new colleague
Name it, write its instructions in plain language, switch on the tools it needs and publish a version.
An animated walk-through of the console: Build: Name it, write its instructions in plain language, switch on the tools it needs and publish a version. Teach: Drop in a handbook. It is parsed, split and indexed on your own servers, and every answer cites the page it came from. Connect: Pick HubSpot, Notion or any of 4,700+ MCP servers, sign in once, and choose which tools the agent may call. Tools that change data can ask first. Chat: Mention an agent, drop in a file and watch every step it takes. A slash runs a command: summarize, explain, regenerate, or one your team wrote. Slack: Mention it in Slack, Mattermost or Telegram. It replies in the thread, with the access of the person who asked. Automate: A routine starts it every Monday, on a new GitHub issue, a Gmail message or a webhook, and keeps the dashboard current. Share: The dashboard gets a stable link and versions. Choose who opens it; every update keeps the link.
Product
One place for the agents your teams actually use.
Chat with them, hand them a task, watch them work, and keep what they make. Real screens from a demo organization, not mock-ups.
01 / 04
Agents that work as a team
A lead agent hands parts of a task to colleagues and writes up what they found. You see who it asked, and what each one said.
How subagents work
What ships today
26 built-in capabilities, switched on per agent.
Each one is a card in the agent’s toolbox: switch it on, set it up, test it in the chat beside it. Plus any tool from a connected MCP server, and capabilities your engineers add in typed Python.
3 capabilities
Web
Search, read pages, or drive a real browser you can take over at a sign-in.
- Web search
- Read web pages
- Web browser
5 capabilities
Files, code and output
Its own computer with files and a shell, and results people can open.
- Calculations
- Sandbox
- Charts
- Image generation
- Apps
7 capabilities
Reasoning and teamwork
Plan, think, hand parts of a job to other agents, and ask you when it is unsure.
- Subagents
- Planning
- Thinking
- Tool search
- Date and time
- Instruction reminders
- Ask the user
4 capabilities
Safety and limits
Redact secrets and personal data, and keep long runs inside their limits.
- Guardrails
- Long conversations
- Media offload
- Tool output limits
1 capability
Chat channels
Look up people and threads in Slack, Telegram and Mattermost, where an @mention runs as the sender.
- Channel lookup
Use cases
Start with one task your team repeats.
29 tutorials give you a sample input, the configuration, the exact prompt and a check you can run. Pick the one whose result somebody on the team can judge.
Documents and knowledge
6 tutorials
Customer support
5 tutorials
Research and analysis
6 tutorials
Automation and teams of agents
5 tutorials
Content and communication
3 tutorials
All 29 tutorials
Starting points with reference runs by the maintainers, not customer case studies.
Integrations
Connect the apps your team already uses.
An agent can read a Notion brief, search SharePoint documents or answer in Slack, under the same access rules and budget. How connections work
Business tools
4,700+ MCP servers in the gallery, and your own. Choose which tools each agent may call.
Documents
Sync Google Drive, S3 or MinIO, Git repositories, websites, SharePoint and OneDrive into knowledge bases.
Events
Start agents on a schedule, a new GitHub issue, a Gmail message or a signed webhook.
Chat tools
Publish an agent to Slack, Mattermost or Telegram, where an @mention runs as the person who sent it.
4,700+
MCP servers in the gallery
Registry listings are publisher metadata; each connection needs its own setup and review.
One agent, eight places it answers
Publish once: the same agent, rules and history everywhere.
- Web chat
- Website widget
- Hosted page
- HTTP API
- WebSocket
- Slack
- Mattermost
- Telegram
AgenticOS is an MCP server too
Work in AgenticOS from Claude Code, Codex or Cursor
Add it with one command. Your coding assistant can list and run agents, read runs and what they cost, and search and fill knowledge bases, within what you allowed. Publishing an agent and touching a credential stay in the console.
Terminal
claude mcp add --transport http agenticos https://<your-deployment>/mcpThe client opens the console's consent page, where you choose the organization and what it may do. No key is pasted anywhere.
Control
Every request passes the same controls.
Roles, budgets, approvals, guardrails and a recorded run, whether the request came from chat, Slack, a schedule or the API. Administrators decide who uses each agent, what it may spend and which actions wait for a person.
A person decides what matters
Sensitive tools wait for approval. The run pauses, the request lands here, and the decision is recorded with who made it.
Spend you can see before it happens
Monthly budgets per agent and organization are checked before each model request, and every run records its tokens and cost.
Roles and groups
Roles define what people may do; groups define who you share with. Company sign-in through OIDC, LDAP or Kerberos.
Budgets
Monthly budgets per agent and organization, checked before each model request.
Approvals
Sensitive tools wait for a person. Every decision is recorded with who made it.
Audit log
Hash-chained per organization, verifiable and exportable.
Vault
Envelope-encrypted credentials with a data key per secret. Values are never shown again.
Sandboxes
Code runs in isolated containers with no network unless needed, and CPU, process and time limits.
Sovereignty
Your servers, your models, your rules.
AgenticOS is Apache-2.0 software you can inspect, modify and operate. AI teammates are arriving in every vendor’s cloud; this is the one you run yourself, on any model, with the source open.
Full lock-in platform
0 of 5
layers you own
- The vendor’s
- The vendor’s
- The vendor’s
- The vendor’s
- The vendor’s
Semi lock-in platform
3 of 5
layers you own
- Yours
- Yours
- Yours
- The vendor’s
- The vendor’s
AgenticOS
5 of 5
layers you own
- Process logic
- Data
- Models
- Codebase
- Infrastructure
Process logic, data, models, codebase and infrastructure. Nothing in AgenticOS needs anyone’s permission to keep running.
Setup 1Self-hosted, with hosted models
AgenticOS, documents, vectors and logs run on your servers. Models come from a provider under your own contract and keys, including Azure, Bedrock and Vertex.
- Your servers
- Your contract
- Your keys
Setup 2Fully local
The same layer with open models through Ollama or vLLM on your hardware, local embeddings and local document parsers. Nothing leaves your network.
- Ollama
- vLLM
- Local parsers
AgenticOS
- Runs
- Your infrastructure
- Source
- Apache-2.0
- Models
- 27 providers, including local ones
OpenAI dots
- Runs
- OpenAI’s cloud
- Source
- Proprietary
- Models
- OpenAI only
Grok Bot
- Runs
- xAI’s cloud
- Source
- Proprietary
- Models
- Grok only
Viktor
- Runs
- Viktor’s cloud
- Source
- Proprietary
- Models
- OpenAI, Anthropic, Google, Kimi
n8n
- Runs
- Self-hosted or n8n Cloud
- Source
- Sustainable Use License
- Models
- Many, including Ollama
| Product | Where it runs | Source | Models |
|---|---|---|---|
| AgenticOS | Your infrastructure | Apache-2.0 | 27 providers, including local ones |
| OpenAI dots | OpenAI’s cloud | Proprietary | OpenAI only |
| Grok Bot | xAI’s cloud | Proprietary | Grok only |
| Viktor | Viktor’s cloud | Proprietary | OpenAI, Anthropic, Google, Kimi |
| n8n | Self-hosted or n8n Cloud | Sustainable Use License | Many, including Ollama |
Each cell comes from the vendor’s own pages. Every comparison, in detail
For engineers
Code defines, configuration composes.
Business teams assemble agents in the browser. Engineers extend what there is to assemble, review it like any other code, and own every line. Add a capability
async def current_weather(city: str) -> Forecast:
"""Get the current weather for a city.
Use this when an answer depends on today's conditions
rather than a seasonal average.
"""
return await weather_api.current(city)
class WeatherConfig(BaseModel):
units: Literal["metric", "imperial"] = Field(
default="metric",
description="metric for °C and km/h, imperial for °F and mph.",
)
@register(
id="weather", # stable in every published spec
name="Weather",
description="Read current conditions instead of assuming them.",
config_schema=WeatherConfig, # the settings form, validated at publish
scopes=("weather:read",), # what a role or an API key needs
)
def build(ctx: CapabilityBuildContext) -> Weather:
return Weather(config=WeatherConfig.model_validate(ctx.config))A capability is typed Python: the tool’s docstring is the prompt the model reads, the config is a validated model the console draws as a form, and scopes decide who may use it.
Code defines what exists
Capabilities, connectors and catalog entries are typed Python on Pydantic AI. Configuration only reaches what code registered.
Agents are files
A spec exports to YAML for your repository, with versions, rollback and a diff that shows what changed.
One runner behind every surface
Chat, Slack, a routine and the API run the same agent, through the same budgets, approvals and recorded run.
An API, a WebSocket and an MCP server
Call agents from your own code, stream tokens, or drive the organization from Claude Code or Cursor.
Built on
- Pydantic AI
- Python
- FastAPI
- PostgreSQL + pgvector
- Redis or Valkey
- Prefect workers
- Next.js
- Docker Compose
Built in the open by 9 contributors
Every change reviewed in public, every release tagged on GitHub.
Pricing
Free to run. Help when you want it.
No seat fees and no credits: AgenticOS is Apache-2.0 software. Run it yourself, or have the team that builds it put it into production for you. Model usage is billed by your own provider, under your own contract. Compare with per-seat products
Run it yourself
$0Apache-2.0 licence
The whole product, on your own servers. You pay for the infrastructure and the model usage you choose.
- Every capability, surface and integration
- Company sign-in, roles, budgets, approvals and audit in every deployment
- Commercial use, modification and private deployment
- Community help on GitHub
With the team that builds it
Rolled out by Vstorm
Per projectImplementation and tailoring
The team that builds AgenticOS sets it up in your environment and puts your first workflows into production.
- Your processes mapped and redesigned for agents
- Agents built and tested in production with your team
- Training, runbooks and documentation at handover
- No subscription, no seat count, no long-term commitment
Run by Vstorm
OngoingIf you are not ready to own it yet
Vstorm maintains the platform for you, on your infrastructure, until your own team takes it over.
- Vstorm maintains the platform for you
- Your team takes it over after training, when it is ready
- Nothing in AgenticOS needs Vstorm’s permission to keep running
- The platform, code, data, infrastructure and models stay yours
Built by
Want it running in your company? We build it in.
The team that builds AgenticOS maps your processes, builds the agents inside your environment and hands the system over to your people, or keeps running it for you.
Talk to the founders
15 or 30 minutes, free of charge
Antoni KozelskiCEO and co-founder
Bartosz Adam GonczarekCo-founder and CTO
You leave with
- A live look at real implementations relevant to your industry.
- One concrete, high-value process from your business mapped to an agentic AI solution.
The first meeting maps which processes AgenticOS should take on first, what already exists for them and what the first ninety days look like. You speak with the engineers who would do the work.
Reply within 48 hours, from a founder or principal engineer. Or write to info@vstorm.co.
Partners and recognition
- PydanticOfficial implementation partnerAgenticOS runs on Pydantic AI
- OpenAISelect PartnerPartner on the models Vstorm ships
- AnthropicClaude Certified Services PartnerPartner on the models Vstorm ships
Agentic AI Foundation memberThe first agentic AI engineering company to join- ClutchTop AI company2023
- DeloitteTechnology Fast 502024
- The ManifestTop AI company2024
How we deliver it in three phases.
A transformation manager runs the change alongside: the roadmap, adoption and upskilling. You pay for implementation and tailoring. No subscription, no seat count, no long-term commitment.
01 · Vstorm engineers lead
Foundation
The open-source base, already built and running in production, set up in your environment.
- The layer running in your environment
- Your first workflow mapped
02 · Built together with your team
Engineering
Forward-deployed engineers map your processes, redesign them for agents, then build and tailor the system in production.
- Processes redesigned for agents
- Agents live on your processes
03 · Your team runs it
Handover
Upskilling, runbooks, documentation and pairing, so your team runs and evolves the system when we step back.
- Your people trained to work with agents
- A team that runs it without Vstorm
Agents we have put in production for clients.
All case studiesFigures from each client’s case study on vstorm.co. These are Vstorm projects, not all of them built on AgenticOS.
- production deployments
- 30+
- AI engineers
- 25+
- building AI systems since
- 2017
- PyPI downloads of our open-source libraries
- 9M+
Quick start
Running in one command.
The installer pulls the published images and starts a deployment with a working agent in it. Everything stays on the machine you run it on.
$ curl - Host
- 4 vCPU · 8 GB RAM
- Needs
- Docker Compose
- Then open
- localhost:3000
- Step 1
Run one command
On macOS or Linux, or inside WSL2 on Windows. It checks Docker first and changes nothing else on the machine.
- Step 2
Answer four questions
A model provider and its key, your login and an organization name. The key is stored encrypted in your own database.
- Step 3
Open the console
A working agent is already there. Ask it something, drop a PDF into a knowledge base, connect a tool.
FAQ
What teams ask before they start.
Licence, local models, control, sign-in and what it takes to run.
Something we did not answer?
Ask the founders. 15 or 30 minutes, free of charge.
Is AgenticOS free for commercial use?
Yes. It is licensed under Apache-2.0, which allows commercial use, modification and private deployment. You pay for the infrastructure you run it on and for the model providers and services you choose.
Is AgenticOS a product or a service?
Both. It is an open-source product you can install and run yourself, and Vstorm tailors it to your processes, data and integrations when you want help. Either way the code, the data and the infrastructure stay yours.
Can Vstorm roll it out for us, and what does that cost?
Yes. Vstorm, the team that builds AgenticOS, maps your processes, builds the agents inside your environment and hands the system over to your team, or maintains it for you. You pay for implementation and tailoring: no subscription, no seat count and no long-term commitment. A first call is 15 or 30 minutes and free of charge.
Do we need engineers to use it?
Not to build agents: instructions, tools and knowledge are set up in the browser by the people who know the work. You need someone to operate the deployment, and engineers only when you want a capability that does not exist yet, which they add in typed Python.
Can it run with local models only?
Yes. Configure Ollama, LiteLLM or an OpenAI-compatible server such as vLLM or LM Studio, use local embedding models through Ollama, and parse documents with PyMuPDF or LiteParse. Check each destination you configure: self-hosting the console alone does not make every model, parser or tool local.
How does it keep agents under control?
Monthly budgets per agent and per organization are checked before each model request. Sensitive tools wait for a person’s approval. Optional guardrails redact secrets and personal data, and every run is recorded with its agent version, tools, tokens and cost.
Which sign-in methods and roles does it support?
Email and password with magic links, Google, OIDC single sign-on (Entra ID, Okta, Keycloak, Auth0, Authentik), LDAP and Kerberos. Six built-in roles combine with department groups and per-resource sharing. MFA comes from your identity provider; there is no native MFA, SAML or SCIM yet.
How is it different from ChatGPT Enterprise, Copilot Studio or n8n?
AgenticOS runs on your infrastructure with any of 27 model providers and has no seat or credit fee of its own. It builds agents for the organization, published to chat, websites and APIs, rather than assistant seats for individual employees. Compared with n8n it starts from the agent rather than a workflow canvas, and many teams use both.
What does it take to run?
A host with 4 vCPU and 8 GB of RAM, Docker Compose and access to a model provider. Plan for someone to operate the deployment and for subject experts who maintain the instructions and documents.
Start now
Put AI to work on infrastructure you control.
Choose one repeated task and the person who will judge the answers. Set it up with the model and access rules you choose.
Or click through the 14-slide introduction