Show HN: Search2o Simplifies AI Agents

Search2o released an agent server that lets clients discover and run agents through two calls, search() and execAgent(), installed via pip install search2o. The company said the agent server is stateless and can run in a cluster, with configuration and profile changes taking effect on the next agent run on every server without a restart, and that its agent server and UI source code are published on GitHub. Search2o also said conversations and memories are encrypted by default, optionally end to end with a key the user holds, and that the system supports four roles (users, developers, administrators and owners) plus service accounts, built-in authentication and single sign-on.

Security & privacy Secrets live in your vault docs/security/secret-vault.html . Conversations and memories are encrypted docs/security/encryption.html , by default or, optionally, end to end docs/security/encryption.html end-to-end-a-key-you-hold . Data privacy docs/security/data-privacy.html lists how each kind of data is handled. pip install search2o search "Book a trip to New York" execAgent convId, Travel desk , "Book a trip to New York" Agents are discovered by search and build on each other in a conversation. Agent server takes care of everything else, so your agent contains only business logic. A client uses just two calls to discover and run hundreds of agents. Authorization: Bearer <token search → agent execAgent → answer No dependencies to install or run in your organization. Four roles docs/system-management/users-and-roles.html : users, developers, administrators and owners. Service accounts docs/system-management/service-accounts.html for bots and scripts. Built-in authentication docs/system-management/authentication.html as well as single sign‑on docs/system-management/single-sign-on.html . Everything that runs in your organization can be built from the source: the agent server https://github.com/Search2o/agent-server and the UI https://github.com/Search2o/ui . Contact us about.html contact for customizations for internal use. An agent refers to enterprise systems by profile name docs/profiles/overview.html . Dev and prod docs/security/multiple-environments.html environment’s profiles point at its own systems, so the same agent runs unchanged. Agent servers are stateless and can be run in a cluster docs/system-management/agent-servers.html . Configuration and profile changes are live on the next agent run docs/runtime/overview.html how-a-change-reaches-every-agent-server on every server, without restart. Integrate at three levels: from Python expressions docs/agent-definition/python-expressions.html inside an agent, through the REST API docs/rest-api/overview.html , and with hooks docs/hooks/index.html that run your code at the server's decision points.