From install to your first search
Create your account
Your license key appears here when you create your account. No email is sent at this step.
Install the agent server
To evaluate Search2o on your own, use your laptop. To share it with your team, use a server your team can reach. Install the agent server with the bundled UI:
$ pip install search2o
Alternatively, install from the source
Clone the agent server and install it from the clone. The clone does not include the UI.
$ git clone https://github.com/Search2o/agent-server.git
$ pip install -e agent-server
Then build the UI, which needs Node.js 22.18 or newer, and copy the build into the agent server's ui directory:
$ git clone https://github.com/Search2o/ui.git
$ cd ui && npm install && npm run build
$ cp -R dist ../agent-server/search2o/ui
On Windows, copy the dist folder to the same directory. Any web server can serve the UI instead; see How the UI runs.
Run the server
$ SEARCH2O_LICENSE_KEY=your-license-key search2o
PS> $env:SEARCH2O_LICENSE_KEY="your-license-key"
PS> search2o
By default the server starts on port 9020. Use --port to use another port. On a shared server, add --host 0.0.0.0 to the command so your team can reach it. In the next step, use the server’s address and port instead of 127.0.0.1:9020. Agent servers covers the address, the port and TLS.
Self-hosted models using Ollama, vLLM or SGLang do not need a vendor API key. For OpenAI, Anthropic or Gemini, before starting the server, also set one or more of OPENAI_API_KEY, ANTHROPIC_API_KEY or GEMINI_API_KEY.
Set up your account
Open http://127.0.0.1:9020/ui and click New user / Forgot password on the sign-in dialog. Enter the same email you used to create your account in step 1. Enter the code you receive, choose a password, and sign in.
The OpenAPI schema is at /openapi.json, Swagger UI is at /docs and ReDoc at /redoc.
Get started
- Choose an LLM profile. Agents call LLMs through an LLM profile. Use an existing profile or create a new one.
- Draft an agent. Open Agents → Drafts and create a draft. Select your LLM profile, then describe the agent you want in Draft with AI.
- Build two or three agents. Each one typically uses one or two of your enterprise systems and handles one use case.
- Validate and publish them. Then add plain-English descriptions so that search can find them.
- Ask a question. Type it in the search box. The matching agent runs and returns an answer.
Your setup is complete. You can now add users and build the other agents your team needs. For help, click the documentation icon at the top of the UI. To contact us, click the message icon .

