ed-agent

Embed Ed. Your app gets a brain. Your users keep their data.

Ed is a chat agent you drop into an application. It sits one layer above the onde inference SDK: onde gives you an engine to load a model into and send messages to, and Ed adds the part every application built on it was writing for itself.

Lifecycle orchestration. Loading a model is a sequence of state changes, not one call. Ed drives Loading -> Ready (or -> Error) and reports each transition.
Notification. onde exposes status by polling. A UI needs to be told. Ed pushes transitions, tool activity, and replies to an EventSink you implement.
Tool calling. The host registers an explicit JSON-schema capability list. Read-only tools run automatically; mutating tools require host approval.
A flattened reply. Inference returns Result<InferenceResult, _>; a frontend wants one shape covering both outcomes. That's ChatReply.

Ed deliberately does not own sessions, persistence, model catalogues, or downloads. Those differ per application and belong to you.

Usage

use ed_agent::{Ed, GgufModelConfig};

let ed = Ed::new();
ed.load(GgufModelConfig::qwen25_1_5b(), None, None).await?;

let reply = ed.send("Summarise this thread.").await;
println!("{}", reply.reply.unwrap_or_default());

Use Ed::with_agent to provide a ToolExecutor, ApprovalHandler, and EventSink. For Tauri applications, ed-agent-tauri provides that implementation plus the commands a webview invokes.

Any load path, same transitions

Ed::load covers GGUF. For anything else — an assigned model from a server, a UQFF build, a load API that doesn't exist yet — hand the call to load_with and Ed reports the transitions around it:

ed.load_with("Onde model", |engine| {
    engine.load_assigned_model(env, app_id, secret, prompt, sampling)
})
.await?;

Ed takes a closure rather than wrapping each path by name on purpose. Those APIs take types that would drag their crates into Ed's public signature and pin every host to Ed's version of them — and at least one of those types has already moved between onde releases. A closure keeps Ed's dependency surface to onde alone.

Platforms

Ed has no cfg(target_os) gates. onde already splits real and fallback implementations internally, so this crate compiles everywhere and surfaces onde's error on a platform it can't run inference on.

Be careful adding platform gates in a host: a gate narrower than onde's own support list silently disables chat on a platform that would have worked.

#ed-agent

#Usage

#Any load path, same transitions

#Platforms

#Copyright

ed-agent

Usage

Any load path, same transitions

Platforms

Copyright