Make Codex and Claude Code subscriptions go further: a local model router that gives routine agent work to economical models and keeps your strongest model for the hard parts.
Rust ai-agentsai-gatewayclaude-codecodexcoding-agentcost-optimizationdeepseekllm-gatewayllm-routerlocal-firstmodel-routingrusttask-routingtauri
mod runtime_support;
use std::time::Duration;
use runtime_support::*;
#[test]
fn real_listener_precommit_disconnect_reaches_next_frozen_candidate() {
let first = NativeProvider::start(vec![ProviderReply::CloseBeforeSemantic]);
let second = NativeProvider::start(vec![ProviderReply::Complete {
status: 200,
error_kind: None,
body: br#"{"id":"fallback-after-precommit","model":"runtime-native","status":"completed","output":[{"type":"message","role":"assistant","content":[{"type":"output_text","text":"ok"}]}],"usage":{"input_tokens":1,"output_tokens":1,"total_tokens":2}}"#,
}]);
let fixture = RuntimeFixture::launch(&[&first, &second], 2);
let response = fixture.request();
assert_eq!(
response.status,
200,
"{}",
String::from_utf8_lossy(&response.body)
);
let document: serde_json::Value = serde_json::from_slice(&response.body).unwrap();
assert_eq!(document["id"], "fallback-after-precommit");
assert_eq!(first.calls(), 1);
assert_eq!(second.calls(), 1);
}
#[test]
fn real_listener_persists_mechanical_failure_before_next_request() {
let first = NativeProvider::start(vec![ProviderReply::CloseBeforeSemantic]);
let second = NativeProvider::start(vec![
ProviderReply::Complete {
status: 200,
error_kind: None,
body: br#"{"id":"fallback-after-disconnect","model":"runtime-native","status":"completed","output":[{"type":"message","role":"assistant","content":[{"type":"output_text","text":"first"}]}],"usage":{"input_tokens":1,"output_tokens":1,"total_tokens":2}}"#,
},
ProviderReply::Complete {
status: 200,
error_kind: None,
body: br#"{"id":"skipped-cooled-binding","model":"runtime-native","status":"completed","output":[{"type":"message","role":"assistant","content":[{"type":"output_text","text":"second"}]}],"usage":{"input_tokens":1,"output_tokens":1,"total_tokens":2}}"#,
},
]);
let fixture = RuntimeFixture::launch(&[&first, &second], 2);
let first_response = fixture.request();
let second_response = fixture.request();
assert_eq!(first_response.status, 200);
assert_eq!(second_response.status, 200);
assert_eq!(
first.calls(),
1,
"the disconnected binding must be cooling down"
);
assert_eq!(second.calls(), 2);
let document: serde_json::Value = serde_json::from_slice(&second_response.body).unwrap();
assert_eq!(document["id"], "skipped-cooled-binding");
}
#[test]
fn fixed_postcommit_disconnect_never_retries_keys_or_enters_another_plan() {
let primary = NativeProvider::start(vec![ProviderReply::StreamThenClose {
body: b"event: response.created\ndata: {\"type\":\"response.created\",\"response\":{\"id\":\"fixed-prefix\",\"model\":\"runtime-native\"}}\n\nevent: response.output_text.delta\ndata: {\"type\":\"response.output_text.delta\",\"output_index\":0,\"content_index\":0,\"delta\":\"fixed-prefix\"}\n\n",
declared_length: 1_024,
}]);
let other_account = NativeProvider::start(vec![ProviderReply::Complete {
status: 200,
error_kind: None,
body: br#"{"id":"forbidden-account"}"#,
}]);
let other_model = NativeProvider::start(vec![ProviderReply::Complete {
status: 200,
error_kind: None,
body: br#"{"id":"forbidden-model"}"#,
}]);
let fixture = RuntimeFixture::launch_fixed(
&[&primary, &other_account, &other_model],
3,
&[2, 1, 1],
None,
);
let response =
fixture.request_body(br#"{"model":"runtime-model","input":"hello","stream":true}"#);
assert_eq!(response.status, 200);
let body = String::from_utf8_lossy(&response.body);
assert!(body.contains("fixed-prefix"));
assert!(!body.contains("forbidden-"));
assert!(!body.contains("response.completed"));
assert_eq!(primary.calls(), 1, "commit closes same-source key retry");
assert_eq!((other_account.calls(), other_model.calls()), (0, 0));
}
#[test]
fn real_listener_postcommit_disconnect_never_creates_another_attempt() {
hiroute_e2e::p0_execution_receipt!(
"runtime.postcommit",
[
"runtime.postcommit_no_retry",
"runtime.postcommit_provider_count"
]
);
let first = NativeProvider::start(vec![ProviderReply::StreamThenClose {
body: b"event: response.created\ndata: {\"type\":\"response.created\",\"response\":{\"id\":\"semantic-prefix\",\"model\":\"runtime-native\"}}\n\nevent: response.output_text.delta\ndata: {\"type\":\"response.output_text.delta\",\"output_index\":0,\"content_index\":0,\"delta\":\"semantic-prefix\"}\n\n",
declared_length: 1_024,
}]);
let second = NativeProvider::start(vec![ProviderReply::Complete {
status: 200,
error_kind: None,
body: br#"{"id":"forbidden-second"}"#,
}]);
let fixture = RuntimeFixture::launch(&[&first, &second], 2);
let response =
fixture.request_body(br#"{"model":"runtime-model","input":"hello","stream":true}"#);
std::thread::sleep(Duration::from_millis(100));
assert_eq!(response.status, 200);
assert!(
response
.body
.windows(b"semantic-prefix".len())
.any(|window| window == b"semantic-prefix")
);
assert_eq!(first.calls(), 1);
assert_eq!(
second.calls(),
0,
"downstream semantic commit must permanently close fallback"
);
}