What is the aixplain OS?

The aixplain OS is the operating system for accountable AI agents you own, across your operations and services. Your people delegate work to Omni, your developers build with Koder, and every agent answers to an owner. This article explains where the aixplain OS runs, what each part does and who it is for, how every agent is governed and where to start.

What the aixplain OS does

From personal AI to organization-wide results: aixplain is the operating system for accountable AI agents you own. Organizations run more of their operations and services on agents, on any model and any server, air-gapped if needed. People delegate work to Omni; developers build with Koder, and every agent answers to an owner.

The aixplain OS is one system for all your teams, agents and assets. It serves your employees, who work with agents in aixplain Desktop and in the web app, and your customers, who reach the agents you deploy through your apps, chat, voice and web. Teams use it across the front office and the back office, for example:

• Customer service: answer customers in their language from your own systems, sort incoming requests and draft follow-ups that wait for your approval.
• Finance: check each invoice against vendor, pricing, budget and policy rules, and send the exceptions to the right reviewer.
• Legal and HR: check contracts against your rules and flag the clauses a lawyer should read, or screen applicants with a scoring model.
• Communications: watch the news sources you pick and turn them into a daily report.
• Engineering: build an agent and ship it as a secure API, or bring the agents you already run under the same controls.

When work repeats, it becomes an agent with an owner. Everything your teams create is kept in your workspace's Library, so each new use case starts from what the last one built.

Who it is for

• AI operators: everyone on your team hands work to Omni in their own words and keeps the judgment.
• Developers: build software and agents with Koder and the Agents SDK, and ship each agent as an API with its own keys and limits.
• Admins: decide who can use each agent, model and tool, what inspectors enforce and what each group and person can spend.

Where the aixplain OS runs

By default, the aixplain OS runs on cloud: we run it for you, and you use it at app.aixplain.com with nothing to install. Omni on the web, the Marketplace and the agents you deploy all run there, 24/7 behind a secure API. See Use the aixplain OS.

Two other places run the same OS:

• Your own servers: on-prem, in your VPC or air-gapped, with the Enterprise plan. See Run the aixplain OS on your own servers.
• Your computer: aixplain Desktop runs Omni and Koder locally, with local or cloud models. See Get started with aixplain Desktop.

The parts of the aixplain OS

Each part has its own article with the details. Start with the one that matches what you want to do.

Omni: delegate work

Everyone in your workspace gets their own Omni, the AI agent you delegate work to. Ask in your own words, and Omni brings in the right tool, skill and agent for each task, asks before it sends, spends, books or connects, and repeats the work on a schedule. When work repeats, it creates an agent you use, schedule and share. See Get started with Omni.

Koder: build software and agents

Koder is the AI your developers build with. Plan, code and test software, and build and evaluate agents that deploy on cloud or on your servers, on any model, including local ones, in the desktop app or the CLI. See Get started with Koder.

Scan: find security issues

Scan reads a repository for security bugs and lists what it finds, worst first, with a trace of how each one is reached and a proposed fix. It also gates CI on the severity you set. See Find security issues with Scan.

Design: make screens, slides and video

Design builds landing pages, dashboards, app screens, slide decks, animated intros and vertical video as files in your project. Edit by pointing at the preview, and export to PDF, PPTX or MP4. See Make screens, slides and video with Design.

Agents and the Agents SDK: build, test and deploy your own agents

Build an agent with Omni from a plain-language description, with Koder or in your own code with the Agents SDK (pip install aixplain). Then test it and deploy it on cloud or your own servers, where it gets a REST API and MCP endpoint, a key, rate limits and a budget. See Build and deploy agents and the Agents SDK docs.

aixplain Desktop: work locally

aixplain Desktop is the aixplain app for your computer. It brings Koder, Scan, Design, Omni and Agents into one app, and you choose which model does the work, including a local one. You can also pair your phone with it to follow and approve its work from anywhere. See Get started with aixplain Desktop.

Chat, Studio, Dashboard and Settings

When you sign in, you use the aixplain OS through its app: Chat to work with Omni, Studio to build agents, the Dashboard to watch runs and usage, and Settings for keys, billing and your profile, all inside your workspace. On cloud, we run the OS for you at app.aixplain.com; on your own servers, it runs at your organization's address. See Use the aixplain OS and Manage your workspace.

The Marketplace: models, tools and agents

The Marketplace holds the models, tools and integrations your agents use, each with its price on its card. Bring your own models and tools too: they stay private in your Library. Share an agent with the workspaces you choose, or publish it to the Marketplace. See Find models, tools and integrations in the Marketplace.

How every agent is governed

Every agent you build on the aixplain OS is a service you own. The tools you connect keep their own access rules, and you add more.

• An owner: every agent has a named owner from the start, so each result traces to a person.
• Inspectors: inspector agents check every input, step and output for personal data, jailbreaks, unsafe content or hallucinations, then edit, rerun, stop or escalate a run that breaks a rule. Create your own with your policies, rules and thresholds.
• Approvals: agent changes, and Omni and Koder actions, wait for your approval. Anything that sends or spends waits for your Allow.
• Budgets and limits: set budgets and rate limits by workspace, key and model, so an agent stops before it overspends.
• An audit trail: every run records who ran it, each tool call and every decision or approval, with the time and cost of the run. Runs export.
• An off switch: turn off any workspace, key or model.
• Access: decide who can use each agent, model and tool, and which connections Omni can make.

Any model, any server

Use any model in the Marketplace, or your own, and swap it on a live agent with no rebuild. Swapping a model keeps the agent's owner, policies, inspectors and run history, and moving an agent to another server keeps its tools and API. Set a fallback for each model, and on your own servers nothing depends on an outside service.

In aixplain Desktop, you can also use a Claude, ChatGPT or Cursor plan you already pay for, your own provider key or a local model. Local models and tools don't count toward your usage.

Workspaces

A workspace holds your agents, tools, skills, files and Library, along with your API keys, credits, policies and members. Everyone has a personal workspace, and you can create or join team workspaces to share agents and other assets. Policies and Library access cascade from the workspace to its groups and members. On your own servers, one deployment holds a workspace for each entity or team. See Manage your workspace.

What runs alongside the OS

A few services sit next to the aixplain OS rather than inside every install. You reach them from wherever the OS runs, and on your own servers you choose which ones to connect.

• The Marketplace: models, tools, integrations and agents to add to your agents, each with its price on its card.
• Hosted models and tools: models and tools aixplain runs for you, used through the Marketplace.
• Your account, credits and billing: sign-in, plans, prepaid credits and top-ups. See Your account, plans and billing.
• Licensing and support: your plan's license and help from our team.

Get started

Pick the place that fits what you want to do first:

• Hand off work: sign up at app.aixplain.com and ask Omni. See Get started with Omni.
• Work in your browser: see Use the aixplain OS.
• Work on your computer: download the app. See Get started with aixplain Desktop.
• Write code: see Get started with Koder.
• Build your own agents: see Build and deploy agents.
• Set up your team: see Manage your workspace.
• Run it on your own servers: see Enterprise or contact us.

Pricing and usage

The Core plan is free, with free credits to use within 7 days and no card needed; after that you pay as you go. One credit is one US dollar. Scale, in early access, adds groups, access and budget policies and usage dashboards for teams on cloud. Enterprise runs the aixplain OS on your own servers and is priced to your deployment. Models and tools you host on your servers use no credits. For plans, see Pricing and Account plans and billing.

Privacy and data

Your data is always yours: we never sell it or train on it, wherever aixplain runs. On cloud, data is encrypted at rest and in transit, and zero data retention models keep nothing you send. On your own servers, your data never leaves them. In aixplain Desktop, a cloud model sees only the step it handles. See Privacy and data and Security.

Frequently asked questions