Build with the engine that powers webAI.
Access the same intelligence platform behind our app through developer-first SDKs. Deploy specialized Personas, document search, and cross-device connectivity on infrastructure you control.
Talk to our teamOne model
to build against.
Everything you build on webAI
follows the same shape.
App
Your product, with the kernel embedded.
Persona
Your domain expert.
Agent
Compute for one conversation.
Turn
Input in, response streamed out.
Kernel
Identity, data, and transport.
App
Your product, with the kernel embedded.
Persona
Your domain expert.
Agent
Compute for one conversation.
Kernel
Identity, data, and transport.
Turn
Input in, response streamed out.
Four SDKs. Compose what you need.
Give every user, device, and Persona a portable identity in your webAI network. Bring your own IdP — OIDC and Microsoft Entra, with SCIM directory sync — and let webAI handle the rest.
Code you can build with
Start the local engine, open a typed client over an encrypted local channel, and stream turns.
// Illustrative — the shape of a webAI SDK client (Rust; Kotlin mirrors it).use webai::Client; #[tokio::main]async fn main() -> anyhow::Result<()> { // Starts the local webAI engine and connects over mTLS on a local socket. let webai = Client::connect().await?; // Reach a Persona — your domain expert — by id. let persona = webai.persona("onboarding-guide").await?; // Submit a turn; stream the response as it's generated. let mut turn = persona.submit("Draft a welcome email for a new hire").await?; while let Some(chunk) = turn.next().await { print!("{}", chunk?.text()); } Ok(())}Built for both developers and AI
One contract for humans and AI.
A single versioned protobuf API over Connect, gRPC, and gRPC Web: use the Rust or Kotlin SDK, or generate a typed client in any language with docs your coding assistant can scaffold against.
webAI as a tool for your agents.
Tool calling and MCP let your agents build on webAI and call into it with Personas, knowledge, and capabilities, exposed as tools.
Runs on your infrastructure.
On-prem, single box, or on-device: your data and intelligence never leave your control.
Secure and sovereign by design.
mTLS over a local socket (no open port), decentralized identity, content-addressed storage, and your IdP via OIDC, Microsoft Entra, and SCIM.
You own the loop.
Embed the kernel for identity, data, and transport; keep your agent loop, tools, and models.
Deploy anywhere.
Whether you need a server fleet, single box, or on-device, you can use the same SDKs and the same contract wherever your data lives.
The webAI Platform
SDKs
AI capabilities you can build into your own software.
Let’s talk about
what you’d build.
Tell us about your environment and what you need it to do.
Talk to our team