wip: grok provider abstraction

rust/crates/api/src/error.rs

@@ -4,7 +4,10 @@ use std::time::Duration;
 
 #[derive(Debug)]
 pub enum ApiError {
-    MissingApiKey,
+    MissingCredentials {
+        provider: &'static str,
+        env_vars: &'static [&'static str],
+    },
     ExpiredOAuthToken,
     Auth(String),
     InvalidApiKeyEnv(VarError),
@@ -30,13 +33,21 @@ pub enum ApiError {
 }
 
 impl ApiError {
+    #[must_use]
+    pub const fn missing_credentials(
+        provider: &'static str,
+        env_vars: &'static [&'static str],
+    ) -> Self {
+        Self::MissingCredentials { provider, env_vars }
+    }
+
     #[must_use]
     pub fn is_retryable(&self) -> bool {
         match self {
             Self::Http(error) => error.is_connect() || error.is_timeout() || error.is_request(),
             Self::Api { retryable, .. } => *retryable,
             Self::RetriesExhausted { last_error, .. } => last_error.is_retryable(),
-            Self::MissingApiKey
+            Self::MissingCredentials { .. }
             | Self::ExpiredOAuthToken
             | Self::Auth(_)
             | Self::InvalidApiKeyEnv(_)
@@ -51,12 +62,11 @@ impl ApiError {
 impl Display for ApiError {
     fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
         match self {
-            Self::MissingApiKey => {
-                write!(
+            Self::MissingCredentials { provider, env_vars } => write!(
                 f,
-                    "ANTHROPIC_AUTH_TOKEN or ANTHROPIC_API_KEY is not set; export one before calling the Anthropic API"
-                )
-            }
+                "missing {provider} credentials; export {} before calling the {provider} API",
+                env_vars.join(" or ")
+            ),
             Self::ExpiredOAuthToken => {
                 write!(
                     f,
@@ -65,10 +75,7 @@ impl Display for ApiError {
             }
             Self::Auth(message) => write!(f, "auth error: {message}"),
             Self::InvalidApiKeyEnv(error) => {
-                write!(
-                    f,
-                    "failed to read ANTHROPIC_AUTH_TOKEN / ANTHROPIC_API_KEY: {error}"
-                )
+                write!(f, "failed to read credential environment variable: {error}")
             }
             Self::Http(error) => write!(f, "http error: {error}"),
             Self::Io(error) => write!(f, "io error: {error}"),
@@ -81,20 +88,14 @@ impl Display for ApiError {
                 ..
             } => match (error_type, message) {
                 (Some(error_type), Some(message)) => {
-                    write!(
-                        f,
-                        "anthropic api returned {status} ({error_type}): {message}"
-                    )
+                    write!(f, "api returned {status} ({error_type}): {message}")
                 }
-                _ => write!(f, "anthropic api returned {status}: {body}"),
+                _ => write!(f, "api returned {status}: {body}"),
             },
             Self::RetriesExhausted {
                 attempts,
                 last_error,
-            } => write!(
-                f,
-                "anthropic api failed after {attempts} attempts: {last_error}"
-            ),
+            } => write!(f, "api failed after {attempts} attempts: {last_error}"),
             Self::InvalidSseFrame(message) => write!(f, "invalid sse frame: {message}"),
             Self::BackoffOverflow {
                 attempt,

rust/crates/api/src/lib.rs

@@ -1,13 +1,19 @@
 mod client;
 mod error;
+mod providers;
 mod sse;
 mod types;
 
 pub use client::{
-    oauth_token_is_expired, read_base_url, resolve_saved_oauth_token, resolve_startup_auth_source,
-    AnthropicClient, AuthSource, MessageStream, OAuthTokenSet,
+    oauth_token_is_expired, read_base_url, read_xai_base_url, resolve_saved_oauth_token,
+    resolve_startup_auth_source, MessageStream, OAuthTokenSet, ProviderClient,
 };
 pub use error::ApiError;
+pub use providers::anthropic::{AnthropicClient, AuthSource};
+pub use providers::openai_compat::{OpenAiCompatClient, OpenAiCompatConfig};
+pub use providers::{
+    detect_provider_kind, max_tokens_for_model, resolve_model_alias, ProviderKind,
+};
 pub use sse::{parse_frame, SseParser};
 pub use types::{
     ContentBlockDelta, ContentBlockDeltaEvent, ContentBlockStartEvent, ContentBlockStopEvent,

rust/crates/api/src/providers/mod.rs

@@ -0,0 +1,216 @@
+use std::future::Future;
+use std::pin::Pin;
+
+use crate::error::ApiError;
+use crate::types::{MessageRequest, MessageResponse};
+
+pub mod anthropic;
+pub mod openai_compat;
+
+pub type ProviderFuture<'a, T> = Pin<Box<dyn Future<Output = Result<T, ApiError>> + Send + 'a>>;
+
+pub trait Provider {
+    type Stream;
+
+    fn send_message<'a>(
+        &'a self,
+        request: &'a MessageRequest,
+    ) -> ProviderFuture<'a, MessageResponse>;
+
+    fn stream_message<'a>(
+        &'a self,
+        request: &'a MessageRequest,
+    ) -> ProviderFuture<'a, Self::Stream>;
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum ProviderKind {
+    Anthropic,
+    Xai,
+    OpenAi,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub struct ProviderMetadata {
+    pub provider: ProviderKind,
+    pub auth_env: &'static str,
+    pub base_url_env: &'static str,
+    pub default_base_url: &'static str,
+}
+
+const MODEL_REGISTRY: &[(&str, ProviderMetadata)] = &[
+    (
+        "opus",
+        ProviderMetadata {
+            provider: ProviderKind::Anthropic,
+            auth_env: "ANTHROPIC_API_KEY",
+            base_url_env: "ANTHROPIC_BASE_URL",
+            default_base_url: anthropic::DEFAULT_BASE_URL,
+        },
+    ),
+    (
+        "sonnet",
+        ProviderMetadata {
+            provider: ProviderKind::Anthropic,
+            auth_env: "ANTHROPIC_API_KEY",
+            base_url_env: "ANTHROPIC_BASE_URL",
+            default_base_url: anthropic::DEFAULT_BASE_URL,
+        },
+    ),
+    (
+        "haiku",
+        ProviderMetadata {
+            provider: ProviderKind::Anthropic,
+            auth_env: "ANTHROPIC_API_KEY",
+            base_url_env: "ANTHROPIC_BASE_URL",
+            default_base_url: anthropic::DEFAULT_BASE_URL,
+        },
+    ),
+    (
+        "grok",
+        ProviderMetadata {
+            provider: ProviderKind::Xai,
+            auth_env: "XAI_API_KEY",
+            base_url_env: "XAI_BASE_URL",
+            default_base_url: openai_compat::DEFAULT_XAI_BASE_URL,
+        },
+    ),
+    (
+        "grok-3",
+        ProviderMetadata {
+            provider: ProviderKind::Xai,
+            auth_env: "XAI_API_KEY",
+            base_url_env: "XAI_BASE_URL",
+            default_base_url: openai_compat::DEFAULT_XAI_BASE_URL,
+        },
+    ),
+    (
+        "grok-mini",
+        ProviderMetadata {
+            provider: ProviderKind::Xai,
+            auth_env: "XAI_API_KEY",
+            base_url_env: "XAI_BASE_URL",
+            default_base_url: openai_compat::DEFAULT_XAI_BASE_URL,
+        },
+    ),
+    (
+        "grok-3-mini",
+        ProviderMetadata {
+            provider: ProviderKind::Xai,
+            auth_env: "XAI_API_KEY",
+            base_url_env: "XAI_BASE_URL",
+            default_base_url: openai_compat::DEFAULT_XAI_BASE_URL,
+        },
+    ),
+    (
+        "grok-2",
+        ProviderMetadata {
+            provider: ProviderKind::Xai,
+            auth_env: "XAI_API_KEY",
+            base_url_env: "XAI_BASE_URL",
+            default_base_url: openai_compat::DEFAULT_XAI_BASE_URL,
+        },
+    ),
+];
+
+#[must_use]
+pub fn resolve_model_alias(model: &str) -> String {
+    let trimmed = model.trim();
+    let lower = trimmed.to_ascii_lowercase();
+    MODEL_REGISTRY
+        .iter()
+        .find_map(|(alias, metadata)| {
+            (*alias == lower).then_some(match metadata.provider {
+                ProviderKind::Anthropic => match *alias {
+                    "opus" => "claude-opus-4-6",
+                    "sonnet" => "claude-sonnet-4-6",
+                    "haiku" => "claude-haiku-4-5-20251213",
+                    _ => trimmed,
+                },
+                ProviderKind::Xai => match *alias {
+                    "grok" | "grok-3" => "grok-3",
+                    "grok-mini" | "grok-3-mini" => "grok-3-mini",
+                    "grok-2" => "grok-2",
+                    _ => trimmed,
+                },
+                ProviderKind::OpenAi => trimmed,
+            })
+        })
+        .map_or_else(|| trimmed.to_string(), ToOwned::to_owned)
+}
+
+#[must_use]
+pub fn metadata_for_model(model: &str) -> Option<ProviderMetadata> {
+    let canonical = resolve_model_alias(model);
+    if canonical.starts_with("claude") {
+        return Some(ProviderMetadata {
+            provider: ProviderKind::Anthropic,
+            auth_env: "ANTHROPIC_API_KEY",
+            base_url_env: "ANTHROPIC_BASE_URL",
+            default_base_url: anthropic::DEFAULT_BASE_URL,
+        });
+    }
+    if canonical.starts_with("grok") {
+        return Some(ProviderMetadata {
+            provider: ProviderKind::Xai,
+            auth_env: "XAI_API_KEY",
+            base_url_env: "XAI_BASE_URL",
+            default_base_url: openai_compat::DEFAULT_XAI_BASE_URL,
+        });
+    }
+    None
+}
+
+#[must_use]
+pub fn detect_provider_kind(model: &str) -> ProviderKind {
+    if let Some(metadata) = metadata_for_model(model) {
+        return metadata.provider;
+    }
+    if anthropic::has_auth_from_env_or_saved().unwrap_or(false) {
+        return ProviderKind::Anthropic;
+    }
+    if openai_compat::has_api_key("OPENAI_API_KEY") {
+        return ProviderKind::OpenAi;
+    }
+    if openai_compat::has_api_key("XAI_API_KEY") {
+        return ProviderKind::Xai;
+    }
+    ProviderKind::Anthropic
+}
+
+#[must_use]
+pub fn max_tokens_for_model(model: &str) -> u32 {
+    let canonical = resolve_model_alias(model);
+    if canonical.contains("opus") {
+        32_000
+    } else {
+        64_000
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::{detect_provider_kind, max_tokens_for_model, resolve_model_alias, ProviderKind};
+
+    #[test]
+    fn resolves_grok_aliases() {
+        assert_eq!(resolve_model_alias("grok"), "grok-3");
+        assert_eq!(resolve_model_alias("grok-mini"), "grok-3-mini");
+        assert_eq!(resolve_model_alias("grok-2"), "grok-2");
+    }
+
+    #[test]
+    fn detects_provider_from_model_name_first() {
+        assert_eq!(detect_provider_kind("grok"), ProviderKind::Xai);
+        assert_eq!(
+            detect_provider_kind("claude-sonnet-4-6"),
+            ProviderKind::Anthropic
+        );
+    }
+
+    #[test]
+    fn keeps_existing_max_token_heuristic() {
+        assert_eq!(max_tokens_for_model("opus"), 32_000);
+        assert_eq!(max_tokens_for_model("grok-3"), 64_000);
+    }
+}

rust/crates/api/tests/client_integration.rs

@@ -3,9 +3,9 @@ use std::sync::Arc;
 use std::time::Duration;
 
 use api::{
-    AnthropicClient, ApiError, ContentBlockDelta, ContentBlockDeltaEvent, ContentBlockStartEvent,
-    InputContentBlock, InputMessage, MessageDeltaEvent, MessageRequest, OutputContentBlock,
-    StreamEvent, ToolChoice, ToolDefinition,
+    AnthropicClient, ApiError, AuthSource, ContentBlockDelta, ContentBlockDeltaEvent,
+    ContentBlockStartEvent, InputContentBlock, InputMessage, MessageDeltaEvent, MessageRequest,
+    OutputContentBlock, ProviderClient, StreamEvent, ToolChoice, ToolDefinition,
 };
 use serde_json::json;
 use tokio::io::{AsyncReadExt, AsyncWriteExt};
@@ -195,6 +195,47 @@ async fn retries_retryable_failures_before_succeeding() {
     assert_eq!(state.lock().await.len(), 2);
 }
 
+#[tokio::test]
+async fn provider_client_dispatches_anthropic_requests() {
+    let state = Arc::new(Mutex::new(Vec::<CapturedRequest>::new()));
+    let server = spawn_server(
+        state.clone(),
+        vec![http_response(
+            "200 OK",
+            "application/json",
+            "{\"id\":\"msg_provider\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"text\",\"text\":\"Dispatched\"}],\"model\":\"claude-3-7-sonnet-latest\",\"stop_reason\":\"end_turn\",\"stop_sequence\":null,\"usage\":{\"input_tokens\":3,\"output_tokens\":2}}",
+        )],
+    )
+    .await;
+
+    let client = ProviderClient::from_model_with_anthropic_auth(
+        "claude-sonnet-4-6",
+        Some(AuthSource::ApiKey("test-key".to_string())),
+    )
+    .expect("anthropic provider client should be constructed");
+    let client = match client {
+        ProviderClient::Anthropic(client) => {
+            ProviderClient::Anthropic(client.with_base_url(server.base_url()))
+        }
+        other => panic!("expected anthropic provider, got {other:?}"),
+    };
+
+    let response = client
+        .send_message(&sample_request(false))
+        .await
+        .expect("provider-dispatched request should succeed");
+
+    assert_eq!(response.total_tokens(), 5);
+
+    let captured = state.lock().await;
+    let request = captured.first().expect("server should capture request");
+    assert_eq!(request.path, "/v1/messages");
+    assert_eq!(
+        request.headers.get("x-api-key").map(String::as_str),
+        Some("test-key")
+    );
+}
+
 #[tokio::test]
 async fn surfaces_retry_exhaustion_for_persistent_retryable_errors() {
     let state = Arc::new(Mutex::new(Vec::<CapturedRequest>::new()));

rust/crates/api/tests/openai_compat_integration.rs

@@ -0,0 +1,415 @@
+use std::collections::HashMap;
+use std::ffi::OsString;
+use std::sync::Arc;
+use std::sync::{Mutex as StdMutex, OnceLock};
+
+use api::{
+    ContentBlockDelta, ContentBlockDeltaEvent, ContentBlockStartEvent, ContentBlockStopEvent,
+    InputContentBlock, InputMessage, MessageRequest, OpenAiCompatClient, OpenAiCompatConfig,
+    OutputContentBlock, ProviderClient, StreamEvent, ToolChoice, ToolDefinition,
+};
+use serde_json::json;
+use tokio::io::{AsyncReadExt, AsyncWriteExt};
+use tokio::net::TcpListener;
+use tokio::sync::Mutex;
+
+#[tokio::test]
+async fn send_message_uses_openai_compatible_endpoint_and_auth() {
+    let state = Arc::new(Mutex::new(Vec::<CapturedRequest>::new()));
+    let body = concat!(
+        "{",
+        "\"id\":\"chatcmpl_test\",",
+        "\"model\":\"grok-3\",",
+        "\"choices\":[{",
+        "\"message\":{\"role\":\"assistant\",\"content\":\"Hello from Grok\",\"tool_calls\":[]},",
+        "\"finish_reason\":\"stop\"",
+        "}],",
+        "\"usage\":{\"prompt_tokens\":11,\"completion_tokens\":5}",
+        "}"
+    );
+    let server = spawn_server(
+        state.clone(),
+        vec![http_response("200 OK", "application/json", body)],
+    )
+    .await;
+
+    let client = OpenAiCompatClient::new("xai-test-key", OpenAiCompatConfig::xai())
+        .with_base_url(server.base_url());
+    let response = client
+        .send_message(&sample_request(false))
+        .await
+        .expect("request should succeed");
+
+    assert_eq!(response.model, "grok-3");
+    assert_eq!(response.total_tokens(), 16);
+    assert_eq!(
+        response.content,
+        vec![OutputContentBlock::Text {
+            text: "Hello from Grok".to_string(),
+        }]
+    );
+
+    let captured = state.lock().await;
+    let request = captured.first().expect("server should capture request");
+    assert_eq!(request.path, "/chat/completions");
+    assert_eq!(
+        request.headers.get("authorization").map(String::as_str),
+        Some("Bearer xai-test-key")
+    );
+    let body: serde_json::Value = serde_json::from_str(&request.body).expect("json body");
+    assert_eq!(body["model"], json!("grok-3"));
+    assert_eq!(body["messages"][0]["role"], json!("system"));
+    assert_eq!(body["tools"][0]["type"], json!("function"));
+}
+
+#[tokio::test]
+async fn send_message_accepts_full_chat_completions_endpoint_override() {
+    let state = Arc::new(Mutex::new(Vec::<CapturedRequest>::new()));
+    let body = concat!(
+        "{",
+        "\"id\":\"chatcmpl_full_endpoint\",",
+        "\"model\":\"grok-3\",",
+        "\"choices\":[{",
+        "\"message\":{\"role\":\"assistant\",\"content\":\"Endpoint override works\",\"tool_calls\":[]},",
+        "\"finish_reason\":\"stop\"",
+        "}],",
+        "\"usage\":{\"prompt_tokens\":7,\"completion_tokens\":3}",
+        "}"
+    );
+    let server = spawn_server(
+        state.clone(),
+        vec![http_response("200 OK", "application/json", body)],
+    )
+    .await;
+
+    let endpoint_url = format!("{}/chat/completions", server.base_url());
+    let client = OpenAiCompatClient::new("xai-test-key", OpenAiCompatConfig::xai())
+        .with_base_url(endpoint_url);
+    let response = client
+        .send_message(&sample_request(false))
+        .await
+        .expect("request should succeed");
+
+    assert_eq!(response.total_tokens(), 10);
+
+    let captured = state.lock().await;
+    let request = captured.first().expect("server should capture request");
+    assert_eq!(request.path, "/chat/completions");
+}
+
+#[tokio::test]
+async fn stream_message_normalizes_text_and_multiple_tool_calls() {
+    let state = Arc::new(Mutex::new(Vec::<CapturedRequest>::new()));
+    let sse = concat!(
+        "data: {\"id\":\"chatcmpl_stream\",\"model\":\"grok-3\",\"choices\":[{\"delta\":{\"content\":\"Hello\"}}]}\n\n",
+        "data: {\"id\":\"chatcmpl_stream\",\"choices\":[{\"delta\":{\"tool_calls\":[{\"index\":0,\"id\":\"call_1\",\"function\":{\"name\":\"weather\",\"arguments\":\"{\\\"city\\\":\\\"Paris\\\"}\"}},{\"index\":1,\"id\":\"call_2\",\"function\":{\"name\":\"clock\",\"arguments\":\"{\\\"zone\\\":\\\"UTC\\\"}\"}}]}}]}\n\n",
+        "data: {\"id\":\"chatcmpl_stream\",\"choices\":[{\"delta\":{},\"finish_reason\":\"tool_calls\"}]}\n\n",
+        "data: [DONE]\n\n"
+    );
+    let server = spawn_server(
+        state.clone(),
+        vec![http_response_with_headers(
+            "200 OK",
+            "text/event-stream",
+            sse,
+            &[("x-request-id", "req_grok_stream")],
+        )],
+    )
+    .await;
+
+    let client = OpenAiCompatClient::new("xai-test-key", OpenAiCompatConfig::xai())
+        .with_base_url(server.base_url());
+    let mut stream = client
+        .stream_message(&sample_request(false))
+        .await
+        .expect("stream should start");
+
+    assert_eq!(stream.request_id(), Some("req_grok_stream"));
+
+    let mut events = Vec::new();
+    while let Some(event) = stream.next_event().await.expect("event should parse") {
+        events.push(event);
+    }
+
+    assert!(matches!(events[0], StreamEvent::MessageStart(_)));
+    assert!(matches!(
+        events[1],
+        StreamEvent::ContentBlockStart(ContentBlockStartEvent {
+            content_block: OutputContentBlock::Text { .. },
+            ..
+        })
+    ));
+    assert!(matches!(
+        events[2],
+        StreamEvent::ContentBlockDelta(ContentBlockDeltaEvent {
+            delta: ContentBlockDelta::TextDelta { .. },
+            ..
+        })
+    ));
+    assert!(matches!(
+        events[3],
+        StreamEvent::ContentBlockStart(ContentBlockStartEvent {
+            index: 1,
+            content_block: OutputContentBlock::ToolUse { .. },
+        })
+    ));
+    assert!(matches!(
+        events[4],
+        StreamEvent::ContentBlockDelta(ContentBlockDeltaEvent {
+            index: 1,
+            delta: ContentBlockDelta::InputJsonDelta { .. },
+        })
+    ));
+    assert!(matches!(
+        events[5],
+        StreamEvent::ContentBlockStart(ContentBlockStartEvent {
+            index: 2,
+            content_block: OutputContentBlock::ToolUse { .. },
+        })
+    ));
+    assert!(matches!(
+        events[6],
+        StreamEvent::ContentBlockDelta(ContentBlockDeltaEvent {
+            index: 2,
+            delta: ContentBlockDelta::InputJsonDelta { .. },
+        })
+    ));
+    assert!(matches!(
+        events[7],
+        StreamEvent::ContentBlockStop(ContentBlockStopEvent { index: 1 })
+    ));
+    assert!(matches!(
+        events[8],
+        StreamEvent::ContentBlockStop(ContentBlockStopEvent { index: 2 })
+    ));
+    assert!(matches!(
+        events[9],
+        StreamEvent::ContentBlockStop(ContentBlockStopEvent { index: 0 })
+    ));
+    assert!(matches!(events[10], StreamEvent::MessageDelta(_)));
+    assert!(matches!(events[11], StreamEvent::MessageStop(_)));
+
+    let captured = state.lock().await;
+    let request = captured.first().expect("captured request");
+    assert_eq!(request.path, "/chat/completions");
+    assert!(request.body.contains("\"stream\":true"));
+}
+
+#[tokio::test]
+async fn provider_client_dispatches_xai_requests_from_env() {
+    let _lock = env_lock();
+    let _api_key = ScopedEnvVar::set("XAI_API_KEY", "xai-test-key");
+
+    let state = Arc::new(Mutex::new(Vec::<CapturedRequest>::new()));
+    let server = spawn_server(
+        state.clone(),
+        vec![http_response(
+            "200 OK",
+            "application/json",
+            "{\"id\":\"chatcmpl_provider\",\"model\":\"grok-3\",\"choices\":[{\"message\":{\"role\":\"assistant\",\"content\":\"Through provider client\",\"tool_calls\":[]},\"finish_reason\":\"stop\"}],\"usage\":{\"prompt_tokens\":9,\"completion_tokens\":4}}",
+        )],
+    )
+    .await;
+    let _base_url = ScopedEnvVar::set("XAI_BASE_URL", server.base_url());
+
+    let client =
+        ProviderClient::from_model("grok").expect("xAI provider client should be constructed");
+    assert!(matches!(client, ProviderClient::Xai(_)));
+
+    let response = client
+        .send_message(&sample_request(false))
+        .await
+        .expect("provider-dispatched request should succeed");
+
+    assert_eq!(response.total_tokens(), 13);
+
+    let captured = state.lock().await;
+    let request = captured.first().expect("captured request");
+    assert_eq!(request.path, "/chat/completions");
+    assert_eq!(
+        request.headers.get("authorization").map(String::as_str),
+        Some("Bearer xai-test-key")
+    );
+}
+
+#[derive(Debug, Clone, PartialEq, Eq)]
+struct CapturedRequest {
+    path: String,
+    headers: HashMap<String, String>,
+    body: String,
+}
+
+struct TestServer {
+    base_url: String,
+    join_handle: tokio::task::JoinHandle<()>,
+}
+
+impl TestServer {
+    fn base_url(&self) -> String {
+        self.base_url.clone()
+    }
+}
+
+impl Drop for TestServer {
+    fn drop(&mut self) {
+        self.join_handle.abort();
+    }
+}
+
+async fn spawn_server(
+    state: Arc<Mutex<Vec<CapturedRequest>>>,
+    responses: Vec<String>,
+) -> TestServer {
+    let listener = TcpListener::bind("127.0.0.1:0")
+        .await
+        .expect("listener should bind");
+    let address = listener.local_addr().expect("listener addr");
+    let join_handle = tokio::spawn(async move {
+        for response in responses {
+            let (mut socket, _) = listener.accept().await.expect("accept");
+            let mut buffer = Vec::new();
+            let mut header_end = None;
+            loop {
+                let mut chunk = [0_u8; 1024];
+                let read = socket.read(&mut chunk).await.expect("read request");
+                if read == 0 {
+                    break;
+                }
+                buffer.extend_from_slice(&chunk[..read]);
+                if let Some(position) = find_header_end(&buffer) {
+                    header_end = Some(position);
+                    break;
+                }
+            }
+
+            let header_end = header_end.expect("headers should exist");
+            let (header_bytes, remaining) = buffer.split_at(header_end);
+            let header_text = String::from_utf8(header_bytes.to_vec()).expect("utf8 headers");
+            let mut lines = header_text.split("\r\n");
+            let request_line = lines.next().expect("request line");
+            let path = request_line
+                .split_whitespace()
+                .nth(1)
+                .expect("path")
+                .to_string();
+            let mut headers = HashMap::new();
+            let mut content_length = 0_usize;
+            for line in lines {
+                if line.is_empty() {
+                    continue;
+                }
+                let (name, value) = line.split_once(':').expect("header");
+                let value = value.trim().to_string();
+                if name.eq_ignore_ascii_case("content-length") {
+                    content_length = value.parse().expect("content length");
+                }
+                headers.insert(name.to_ascii_lowercase(), value);
+            }
+
+            let mut body = remaining[4..].to_vec();
+            while body.len() < content_length {
+                let mut chunk = vec![0_u8; content_length - body.len()];
+                let read = socket.read(&mut chunk).await.expect("read body");
+                if read == 0 {
+                    break;
+                }
+                body.extend_from_slice(&chunk[..read]);
+            }
+
+            state.lock().await.push(CapturedRequest {
+                path,
+                headers,
+                body: String::from_utf8(body).expect("utf8 body"),
+            });
+
+            socket
+                .write_all(response.as_bytes())
+                .await
+                .expect("write response");
+        }
+    });
+
+    TestServer {
+        base_url: format!("http://{address}"),
+        join_handle,
+    }
+}
+
+fn find_header_end(bytes: &[u8]) -> Option<usize> {
+    bytes.windows(4).position(|window| window == b"\r\n\r\n")
+}
+
+fn http_response(status: &str, content_type: &str, body: &str) -> String {
+    http_response_with_headers(status, content_type, body, &[])
+}
+
+fn http_response_with_headers(
+    status: &str,
+    content_type: &str,
+    body: &str,
+    headers: &[(&str, &str)],
+) -> String {
+    let mut extra_headers = String::new();
+    for (name, value) in headers {
+        use std::fmt::Write as _;
+        write!(&mut extra_headers, "{name}: {value}\r\n").expect("header write");
+    }
+    format!(
+        "HTTP/1.1 {status}\r\ncontent-type: {content_type}\r\n{extra_headers}content-length: {}\r\nconnection: close\r\n\r\n{body}",
+        body.len()
+    )
+}
+
+fn sample_request(stream: bool) -> MessageRequest {
+    MessageRequest {
+        model: "grok-3".to_string(),
+        max_tokens: 64,
+        messages: vec![InputMessage {
+            role: "user".to_string(),
+            content: vec![InputContentBlock::Text {
+                text: "Say hello".to_string(),
+            }],
+        }],
+        system: Some("Use tools when needed".to_string()),
+        tools: Some(vec![ToolDefinition {
+            name: "weather".to_string(),
+            description: Some("Fetches weather".to_string()),
+            input_schema: json!({
+                "type": "object",
+                "properties": {"city": {"type": "string"}},
+                "required": ["city"]
+            }),
+        }]),
+        tool_choice: Some(ToolChoice::Auto),
+        stream,
+    }
+}
+
+fn env_lock() -> std::sync::MutexGuard<'static, ()> {
+    static LOCK: OnceLock<StdMutex<()>> = OnceLock::new();
+    LOCK.get_or_init(|| StdMutex::new(()))
+        .lock()
+        .unwrap_or_else(|poisoned| poisoned.into_inner())
+}
+
+struct ScopedEnvVar {
+    key: &'static str,
+    previous: Option<OsString>,
+}
+
+impl ScopedEnvVar {
+    fn set(key: &'static str, value: impl AsRef<std::ffi::OsStr>) -> Self {
+        let previous = std::env::var_os(key);
+        std::env::set_var(key, value);
+        Self { key, previous }
+    }
+}
+
+impl Drop for ScopedEnvVar {
+    fn drop(&mut self) {
+        match &self.previous {
+            Some(value) => std::env::set_var(self.key, value),
+            None => std::env::remove_var(self.key),
+        }
+    }
+}

rust/crates/api/tests/provider_client_integration.rs

@@ -0,0 +1,86 @@
+use std::ffi::OsString;
+use std::sync::{Mutex, OnceLock};
+
+use api::{read_xai_base_url, ApiError, AuthSource, ProviderClient, ProviderKind};
+
+#[test]
+fn provider_client_routes_grok_aliases_through_xai() {
+    let _lock = env_lock();
+    let _xai_api_key = EnvVarGuard::set("XAI_API_KEY", Some("xai-test-key"));
+
+    let client = ProviderClient::from_model("grok-mini").expect("grok alias should resolve");
+
+    assert_eq!(client.provider_kind(), ProviderKind::Xai);
+}
+
+#[test]
+fn provider_client_reports_missing_xai_credentials_for_grok_models() {
+    let _lock = env_lock();
+    let _xai_api_key = EnvVarGuard::set("XAI_API_KEY", None);
+
+    let error = ProviderClient::from_model("grok-3")
+        .expect_err("grok requests without XAI_API_KEY should fail fast");
+
+    match error {
+        ApiError::MissingCredentials { provider, env_vars } => {
+            assert_eq!(provider, "xAI");
+            assert_eq!(env_vars, &["XAI_API_KEY"]);
+        }
+        other => panic!("expected missing xAI credentials, got {other:?}"),
+    }
+}
+
+#[test]
+fn provider_client_uses_explicit_anthropic_auth_without_env_lookup() {
+    let _lock = env_lock();
+    let _anthropic_api_key = EnvVarGuard::set("ANTHROPIC_API_KEY", None);
+    let _anthropic_auth_token = EnvVarGuard::set("ANTHROPIC_AUTH_TOKEN", None);
+
+    let client = ProviderClient::from_model_with_anthropic_auth(
+        "claude-sonnet-4-6",
+        Some(AuthSource::ApiKey("anthropic-test-key".to_string())),
+    )
+    .expect("explicit anthropic auth should avoid env lookup");
+
+    assert_eq!(client.provider_kind(), ProviderKind::Anthropic);
+}
+
+#[test]
+fn read_xai_base_url_prefers_env_override() {
+    let _lock = env_lock();
+    let _xai_base_url = EnvVarGuard::set("XAI_BASE_URL", Some("https://example.xai.test/v1"));
+
+    assert_eq!(read_xai_base_url(), "https://example.xai.test/v1");
+}
+
+fn env_lock() -> std::sync::MutexGuard<'static, ()> {
+    static LOCK: OnceLock<Mutex<()>> = OnceLock::new();
+    LOCK.get_or_init(|| Mutex::new(()))
+        .lock()
+        .unwrap_or_else(|poisoned| poisoned.into_inner())
+}
+
+struct EnvVarGuard {
+    key: &'static str,
+    original: Option<OsString>,
+}
+
+impl EnvVarGuard {
+    fn set(key: &'static str, value: Option<&str>) -> Self {
+        let original = std::env::var_os(key);
+        match value {
+            Some(value) => std::env::set_var(key, value),
+            None => std::env::remove_var(key),
+        }
+        Self { key, original }
+    }
+}
+
+impl Drop for EnvVarGuard {
+    fn drop(&mut self) {
+        match &self.original {
+            Some(value) => std::env::set_var(self.key, value),
+            None => std::env::remove_var(self.key),
+        }
+    }
+}

rust/crates/runtime/src/conversation.rs

@@ -118,7 +118,7 @@ where
             tool_executor,
             permission_policy,
             system_prompt,
-            RuntimeFeatureConfig::default(),
+            &RuntimeFeatureConfig::default(),
         )
     }
 
@@ -129,7 +129,7 @@ where
         tool_executor: T,
         permission_policy: PermissionPolicy,
         system_prompt: Vec<String>,
-        feature_config: RuntimeFeatureConfig,
+        feature_config: &RuntimeFeatureConfig,
     ) -> Self {
         let usage_tracker = UsageTracker::from_session(&session);
         Self {
@@ -140,7 +140,7 @@ where
             system_prompt,
             max_iterations: usize::MAX,
             usage_tracker,
-            hook_runner: HookRunner::from_feature_config(&feature_config),
+            hook_runner: HookRunner::from_feature_config(feature_config),
         }
     }
 
@@ -609,7 +609,7 @@ mod tests {
             }),
             PermissionPolicy::new(PermissionMode::DangerFullAccess),
             vec!["system".to_string()],
-            RuntimeFeatureConfig::default().with_hooks(RuntimeHookConfig::new(
+            &RuntimeFeatureConfig::default().with_hooks(RuntimeHookConfig::new(
                 vec![shell_snippet("printf 'blocked by hook'; exit 2")],
                 Vec::new(),
             )),
@@ -675,7 +675,7 @@ mod tests {
             StaticToolExecutor::new().register("add", |_input| Ok("4".to_string())),
             PermissionPolicy::new(PermissionMode::DangerFullAccess),
             vec!["system".to_string()],
-            RuntimeFeatureConfig::default().with_hooks(RuntimeHookConfig::new(
+            &RuntimeFeatureConfig::default().with_hooks(RuntimeHookConfig::new(
                 vec![shell_snippet("printf 'pre hook ran'")],
                 vec![shell_snippet("printf 'post hook ran'")],
             )),
@@ -697,7 +697,7 @@ mod tests {
             "post hook should preserve non-error result: {output:?}"
         );
         assert!(
-            output.contains("4"),
+            output.contains('4'),
             "tool output missing value: {output:?}"
         );
         assert!(

rust/crates/runtime/src/hooks.rs

@@ -51,6 +51,16 @@ pub struct HookRunner {
     config: RuntimeHookConfig,
 }
 
+#[derive(Debug, Clone, Copy)]
+struct HookCommandRequest<'a> {
+    event: HookEvent,
+    tool_name: &'a str,
+    tool_input: &'a str,
+    tool_output: Option<&'a str>,
+    is_error: bool,
+    payload: &'a str,
+}
+
 impl HookRunner {
     #[must_use]
     pub fn new(config: RuntimeHookConfig) -> Self {
@@ -118,14 +128,16 @@ impl HookRunner {
         let mut messages = Vec::new();
 
         for command in commands {
-            match self.run_command(
+            match Self::run_command(
                 command,
+                HookCommandRequest {
                     event,
                     tool_name,
                     tool_input,
                     tool_output,
                     is_error,
-                &payload,
+                    payload: &payload,
+                },
             ) {
                 HookCommandOutcome::Allow { message } => {
                     if let Some(message) = message {
@@ -149,29 +161,23 @@ impl HookRunner {
         HookRunResult::allow(messages)
     }
 
-    fn run_command(
-        &self,
-        command: &str,
-        event: HookEvent,
-        tool_name: &str,
-        tool_input: &str,
-        tool_output: Option<&str>,
-        is_error: bool,
-        payload: &str,
-    ) -> HookCommandOutcome {
+    fn run_command(command: &str, request: HookCommandRequest<'_>) -> HookCommandOutcome {
         let mut child = shell_command(command);
         child.stdin(std::process::Stdio::piped());
         child.stdout(std::process::Stdio::piped());
         child.stderr(std::process::Stdio::piped());
-        child.env("HOOK_EVENT", event.as_str());
-        child.env("HOOK_TOOL_NAME", tool_name);
-        child.env("HOOK_TOOL_INPUT", tool_input);
-        child.env("HOOK_TOOL_IS_ERROR", if is_error { "1" } else { "0" });
-        if let Some(tool_output) = tool_output {
+        child.env("HOOK_EVENT", request.event.as_str());
+        child.env("HOOK_TOOL_NAME", request.tool_name);
+        child.env("HOOK_TOOL_INPUT", request.tool_input);
+        child.env(
+            "HOOK_TOOL_IS_ERROR",
+            if request.is_error { "1" } else { "0" },
+        );
+        if let Some(tool_output) = request.tool_output {
             child.env("HOOK_TOOL_OUTPUT", tool_output);
         }
 
-        match child.output_with_stdin(payload.as_bytes()) {
+        match child.output_with_stdin(request.payload.as_bytes()) {
             Ok(output) => {
                 let stdout = String::from_utf8_lossy(&output.stdout).trim().to_string();
                 let stderr = String::from_utf8_lossy(&output.stderr).trim().to_string();
@@ -189,16 +195,18 @@ impl HookRunner {
                     },
                     None => HookCommandOutcome::Warn {
                         message: format!(
-                            "{} hook `{command}` terminated by signal while handling `{tool_name}`",
-                            event.as_str()
+                            "{} hook `{command}` terminated by signal while handling `{}`",
+                            request.event.as_str(),
+                            request.tool_name
                         ),
                     },
                 }
             }
             Err(error) => HookCommandOutcome::Warn {
                 message: format!(
-                    "{} hook `{command}` failed to start for `{tool_name}`: {error}",
-                    event.as_str()
+                    "{} hook `{command}` failed to start for `{}`: {error}",
+                    request.event.as_str(),
+                    request.tool_name
                 ),
             },
         }

rust/crates/rusty-claude-cli/src/main.rs

@@ -12,8 +12,9 @@ use std::process::Command;
 use std::time::{SystemTime, UNIX_EPOCH};
 
 use api::{
-    resolve_startup_auth_source, AnthropicClient, AuthSource, ContentBlockDelta, InputContentBlock,
-    InputMessage, MessageRequest, MessageResponse, OutputContentBlock,
+    detect_provider_kind, max_tokens_for_model, resolve_model_alias, resolve_startup_auth_source,
+    AnthropicClient, AuthSource, ContentBlockDelta, InputContentBlock, InputMessage,
+    MessageRequest, MessageResponse, OutputContentBlock, ProviderClient, ProviderKind,
     StreamEvent as ApiStreamEvent, ToolChoice, ToolDefinition, ToolResultContentBlock,
 };
 
@@ -35,13 +36,6 @@ use serde_json::json;
 use tools::{execute_tool, mvp_tool_specs, ToolSpec};
 
 const DEFAULT_MODEL: &str = "claude-opus-4-6";
-fn max_tokens_for_model(model: &str) -> u32 {
-    if model.contains("opus") {
-        32_000
-    } else {
-        64_000
-    }
-}
 const DEFAULT_DATE: &str = "2026-03-31";
 const DEFAULT_OAUTH_CALLBACK_PORT: u16 = 4545;
 const VERSION: &str = env!("CARGO_PKG_VERSION");
@@ -288,15 +282,6 @@ fn parse_args(args: &[String]) -> Result<CliAction, String> {
     }
 }
 
-fn resolve_model_alias(model: &str) -> &str {
-    match model {
-        "opus" => "claude-opus-4-6",
-        "sonnet" => "claude-sonnet-4-6",
-        "haiku" => "claude-haiku-4-5-20251213",
-        _ => model,
-    }
-}
-
 fn normalize_allowed_tools(values: &[String]) -> Result<Option<AllowedToolSet>, String> {
     if values.is_empty() {
         return Ok(None);
@@ -980,7 +965,7 @@ struct LiveCli {
     allowed_tools: Option<AllowedToolSet>,
     permission_mode: PermissionMode,
     system_prompt: Vec<String>,
-    runtime: ConversationRuntime<AnthropicRuntimeClient, CliToolExecutor>,
+    runtime: ConversationRuntime<ProviderRuntimeClient, CliToolExecutor>,
     session: SessionHandle,
 }
 
@@ -1920,15 +1905,16 @@ fn build_runtime(
     emit_output: bool,
     allowed_tools: Option<AllowedToolSet>,
     permission_mode: PermissionMode,
-) -> Result<ConversationRuntime<AnthropicRuntimeClient, CliToolExecutor>, Box<dyn std::error::Error>>
+) -> Result<ConversationRuntime<ProviderRuntimeClient, CliToolExecutor>, Box<dyn std::error::Error>>
 {
+    let feature_config = build_runtime_feature_config()?;
     Ok(ConversationRuntime::new_with_features(
         session,
-        AnthropicRuntimeClient::new(model, enable_tools, emit_output, allowed_tools.clone())?,
+        ProviderRuntimeClient::new(model, enable_tools, emit_output, allowed_tools.clone())?,
         CliToolExecutor::new(allowed_tools, emit_output),
         permission_policy(permission_mode),
         system_prompt,
-        build_runtime_feature_config()?,
+        &feature_config,
     ))
 }
 
@@ -1978,26 +1964,33 @@ impl runtime::PermissionPrompter for CliPermissionPrompter {
     }
 }
 
-struct AnthropicRuntimeClient {
+struct ProviderRuntimeClient {
     runtime: tokio::runtime::Runtime,
-    client: AnthropicClient,
+    client: ProviderClient,
     model: String,
     enable_tools: bool,
     emit_output: bool,
     allowed_tools: Option<AllowedToolSet>,
 }
 
-impl AnthropicRuntimeClient {
+impl ProviderRuntimeClient {
     fn new(
         model: String,
         enable_tools: bool,
         emit_output: bool,
         allowed_tools: Option<AllowedToolSet>,
     ) -> Result<Self, Box<dyn std::error::Error>> {
+        let model = resolve_model_alias(&model).to_string();
+        let client = match detect_provider_kind(&model) {
+            ProviderKind::Anthropic => ProviderClient::from_model_with_anthropic_auth(
+                &model,
+                Some(resolve_cli_auth_source()?),
+            )?,
+            ProviderKind::Xai | ProviderKind::OpenAi => ProviderClient::from_model(&model)?,
+        };
         Ok(Self {
             runtime: tokio::runtime::Runtime::new()?,
-            client: AnthropicClient::from_auth(resolve_cli_auth_source()?)
-                .with_base_url(api::read_base_url()),
+            client,
             model,
             enable_tools,
             emit_output,
@@ -2016,7 +2009,7 @@ fn resolve_cli_auth_source() -> Result<AuthSource, Box<dyn std::error::Error>> {
     })?)
 }
 
-impl ApiClient for AnthropicRuntimeClient {
+impl ApiClient for ProviderRuntimeClient {
     #[allow(clippy::too_many_lines)]
     fn stream(&mut self, request: ApiRequest) -> Result<Vec<AssistantEvent>, RuntimeError> {
         let message_request = MessageRequest {
@@ -2054,7 +2047,7 @@ impl ApiClient for AnthropicRuntimeClient {
             let renderer = TerminalRenderer::new();
             let mut markdown_stream = MarkdownStreamState::default();
             let mut events = Vec::new();
-            let mut pending_tool: Option<(String, String, String)> = None;
+            let mut pending_tools: BTreeMap<u32, (String, String, String)> = BTreeMap::new();
             let mut saw_stop = false;
 
             while let Some(event) = stream
@@ -2065,15 +2058,23 @@ impl ApiClient for AnthropicRuntimeClient {
                 match event {
                     ApiStreamEvent::MessageStart(start) => {
                         for block in start.message.content {
-                            push_output_block(block, out, &mut events, &mut pending_tool, true)?;
+                            push_output_block(
+                                block,
+                                0,
+                                out,
+                                &mut events,
+                                &mut pending_tools,
+                                true,
+                            )?;
                         }
                     }
                     ApiStreamEvent::ContentBlockStart(start) => {
                         push_output_block(
                             start.content_block,
+                            start.index,
                             out,
                             &mut events,
-                            &mut pending_tool,
+                            &mut pending_tools,
                             true,
                         )?;
                     }
@@ -2089,18 +2090,18 @@ impl ApiClient for AnthropicRuntimeClient {
                             }
                         }
                         ContentBlockDelta::InputJsonDelta { partial_json } => {
-                            if let Some((_, _, input)) = &mut pending_tool {
+                            if let Some((_, _, input)) = pending_tools.get_mut(&delta.index) {
                                 input.push_str(&partial_json);
                             }
                         }
                     },
-                    ApiStreamEvent::ContentBlockStop(_) => {
+                    ApiStreamEvent::ContentBlockStop(stop) => {
                         if let Some(rendered) = markdown_stream.flush(&renderer) {
                             write!(out, "{rendered}")
                                 .and_then(|()| out.flush())
                                 .map_err(|error| RuntimeError::new(error.to_string()))?;
                         }
-                        if let Some((id, name, input)) = pending_tool.take() {
+                        if let Some((id, name, input)) = pending_tools.remove(&stop.index) {
                             // Display tool call now that input is fully accumulated
                             writeln!(out, "\n{}", format_tool_call_start(&name, &input))
                                 .and_then(|()| out.flush())
@@ -2564,9 +2565,10 @@ fn truncate_for_summary(value: &str, limit: usize) -> String {
 
 fn push_output_block(
     block: OutputContentBlock,
+    block_index: u32,
     out: &mut (impl Write + ?Sized),
     events: &mut Vec<AssistantEvent>,
-    pending_tool: &mut Option<(String, String, String)>,
+    pending_tools: &mut BTreeMap<u32, (String, String, String)>,
     streaming_tool_input: bool,
 ) -> Result<(), RuntimeError> {
     match block {
@@ -2591,7 +2593,7 @@ fn push_output_block(
             } else {
                 input.to_string()
             };
-            *pending_tool = Some((id, name, initial_input));
+            pending_tools.insert(block_index, (id, name, initial_input));
         }
     }
     Ok(())
@@ -2602,11 +2604,13 @@ fn response_to_events(
     out: &mut (impl Write + ?Sized),
 ) -> Result<Vec<AssistantEvent>, RuntimeError> {
     let mut events = Vec::new();
-    let mut pending_tool = None;
+    let mut pending_tools = BTreeMap::new();
 
-    for block in response.content {
-        push_output_block(block, out, &mut events, &mut pending_tool, false)?;
-        if let Some((id, name, input)) = pending_tool.take() {
+    for (index, block) in response.content.into_iter().enumerate() {
+        let index =
+            u32::try_from(index).map_err(|_| RuntimeError::new("response block index overflow"))?;
+        push_output_block(block, index, out, &mut events, &mut pending_tools, false)?;
+        if let Some((id, name, input)) = pending_tools.remove(&index) {
             events.push(AssistantEvent::ToolUse { id, name, input });
         }
     }
@@ -2832,6 +2836,7 @@ mod tests {
     use api::{MessageResponse, OutputContentBlock, Usage};
     use runtime::{AssistantEvent, ContentBlock, ConversationMessage, MessageRole, PermissionMode};
     use serde_json::json;
+    use std::collections::BTreeMap;
     use std::path::PathBuf;
 
     #[test]
@@ -2911,6 +2916,9 @@ mod tests {
         assert_eq!(resolve_model_alias("opus"), "claude-opus-4-6");
         assert_eq!(resolve_model_alias("sonnet"), "claude-sonnet-4-6");
         assert_eq!(resolve_model_alias("haiku"), "claude-haiku-4-5-20251213");
+        assert_eq!(resolve_model_alias("grok"), "grok-3");
+        assert_eq!(resolve_model_alias("grok-mini"), "grok-3-mini");
+        assert_eq!(resolve_model_alias("grok-2"), "grok-2");
         assert_eq!(resolve_model_alias("claude-opus"), "claude-opus");
     }
 
@@ -3378,15 +3386,16 @@ mod tests {
     fn push_output_block_renders_markdown_text() {
         let mut out = Vec::new();
         let mut events = Vec::new();
-        let mut pending_tool = None;
+        let mut pending_tools = BTreeMap::new();
 
         push_output_block(
             OutputContentBlock::Text {
                 text: "# Heading".to_string(),
             },
+            0,
             &mut out,
             &mut events,
-            &mut pending_tool,
+            &mut pending_tools,
             false,
         )
         .expect("text block should render");
@@ -3400,7 +3409,7 @@ mod tests {
     fn push_output_block_skips_empty_object_prefix_for_tool_streams() {
         let mut out = Vec::new();
         let mut events = Vec::new();
-        let mut pending_tool = None;
+        let mut pending_tools = BTreeMap::new();
 
         push_output_block(
             OutputContentBlock::ToolUse {
@@ -3408,20 +3417,83 @@ mod tests {
                 name: "read_file".to_string(),
                 input: json!({}),
             },
+            1,
             &mut out,
             &mut events,
-            &mut pending_tool,
+            &mut pending_tools,
             true,
         )
         .expect("tool block should accumulate");
 
         assert!(events.is_empty());
         assert_eq!(
-            pending_tool,
+            pending_tools.remove(&1),
             Some(("tool-1".to_string(), "read_file".to_string(), String::new(),))
         );
     }
 
+    #[test]
+    fn pending_tools_preserve_multiple_streaming_tool_calls_by_index() {
+        let mut out = Vec::new();
+        let mut events = Vec::new();
+        let mut pending_tools = BTreeMap::new();
+
+        push_output_block(
+            OutputContentBlock::ToolUse {
+                id: "tool-1".to_string(),
+                name: "read_file".to_string(),
+                input: json!({}),
+            },
+            1,
+            &mut out,
+            &mut events,
+            &mut pending_tools,
+            true,
+        )
+        .expect("first tool should accumulate");
+        push_output_block(
+            OutputContentBlock::ToolUse {
+                id: "tool-2".to_string(),
+                name: "grep_search".to_string(),
+                input: json!({}),
+            },
+            2,
+            &mut out,
+            &mut events,
+            &mut pending_tools,
+            true,
+        )
+        .expect("second tool should accumulate");
+
+        pending_tools
+            .get_mut(&1)
+            .expect("first tool pending")
+            .2
+            .push_str("{\"path\":\"src/main.rs\"}");
+        pending_tools
+            .get_mut(&2)
+            .expect("second tool pending")
+            .2
+            .push_str("{\"pattern\":\"TODO\"}");
+
+        assert_eq!(
+            pending_tools.remove(&1),
+            Some((
+                "tool-1".to_string(),
+                "read_file".to_string(),
+                "{\"path\":\"src/main.rs\"}".to_string(),
+            ))
+        );
+        assert_eq!(
+            pending_tools.remove(&2),
+            Some((
+                "tool-2".to_string(),
+                "grep_search".to_string(),
+                "{\"pattern\":\"TODO\"}".to_string(),
+            ))
+        );
+    }
+
     #[test]
     fn response_to_events_preserves_empty_object_json_input_outside_streaming() {
         let mut out = Vec::new();

rust/crates/tools/src/lib.rs

@@ -4,9 +4,9 @@ use std::process::Command;
 use std::time::{Duration, Instant};
 
 use api::{
-    read_base_url, AnthropicClient, ContentBlockDelta, InputContentBlock, InputMessage,
-    MessageRequest, MessageResponse, OutputContentBlock, StreamEvent as ApiStreamEvent, ToolChoice,
-    ToolDefinition, ToolResultContentBlock,
+    max_tokens_for_model, resolve_model_alias, ContentBlockDelta, InputContentBlock, InputMessage,
+    MessageRequest, MessageResponse, OutputContentBlock, ProviderClient,
+    StreamEvent as ApiStreamEvent, ToolChoice, ToolDefinition, ToolResultContentBlock,
 };
 use reqwest::blocking::Client;
 use runtime::{
@@ -1459,14 +1459,14 @@ fn run_agent_job(job: &AgentJob) -> Result<(), String> {
 
 fn build_agent_runtime(
     job: &AgentJob,
-) -> Result<ConversationRuntime<AnthropicRuntimeClient, SubagentToolExecutor>, String> {
+) -> Result<ConversationRuntime<ProviderRuntimeClient, SubagentToolExecutor>, String> {
     let model = job
         .manifest
         .model
         .clone()
         .unwrap_or_else(|| DEFAULT_AGENT_MODEL.to_string());
     let allowed_tools = job.allowed_tools.clone();
-    let api_client = AnthropicRuntimeClient::new(model, allowed_tools.clone())?;
+    let api_client = ProviderRuntimeClient::new(model, allowed_tools.clone())?;
     let tool_executor = SubagentToolExecutor::new(allowed_tools);
     Ok(ConversationRuntime::new(
         Session::new(),
@@ -1635,18 +1635,17 @@ fn format_agent_terminal_output(status: &str, result: Option<&str>, error: Optio
     sections.join("")
 }
 
-struct AnthropicRuntimeClient {
+struct ProviderRuntimeClient {
     runtime: tokio::runtime::Runtime,
-    client: AnthropicClient,
+    client: ProviderClient,
     model: String,
     allowed_tools: BTreeSet<String>,
 }
 
-impl AnthropicRuntimeClient {
+impl ProviderRuntimeClient {
     fn new(model: String, allowed_tools: BTreeSet<String>) -> Result<Self, String> {
-        let client = AnthropicClient::from_env()
-            .map_err(|error| error.to_string())?
-            .with_base_url(read_base_url());
+        let model = resolve_model_alias(&model).to_string();
+        let client = ProviderClient::from_model(&model).map_err(|error| error.to_string())?;
         Ok(Self {
             runtime: tokio::runtime::Runtime::new().map_err(|error| error.to_string())?,
             client,
@@ -1656,7 +1655,7 @@ impl AnthropicRuntimeClient {
     }
 }
 
-impl ApiClient for AnthropicRuntimeClient {
+impl ApiClient for ProviderRuntimeClient {
     fn stream(&mut self, request: ApiRequest) -> Result<Vec<AssistantEvent>, RuntimeError> {
         let tools = tool_specs_for_allowed_tools(Some(&self.allowed_tools))
             .into_iter()
@@ -1668,7 +1667,7 @@ impl ApiClient for AnthropicRuntimeClient {
             .collect::<Vec<_>>();
         let message_request = MessageRequest {
             model: self.model.clone(),
-            max_tokens: 32_000,
+            max_tokens: max_tokens_for_model(&self.model),
             messages: convert_messages(&request.messages),
             system: (!request.system_prompt.is_empty()).then(|| request.system_prompt.join("\n\n")),
             tools: (!tools.is_empty()).then_some(tools),
@@ -1683,7 +1682,7 @@ impl ApiClient for AnthropicRuntimeClient {
                 .await
                 .map_err(|error| RuntimeError::new(error.to_string()))?;
             let mut events = Vec::new();
-            let mut pending_tool: Option<(String, String, String)> = None;
+            let mut pending_tools: BTreeMap<u32, (String, String, String)> = BTreeMap::new();
             let mut saw_stop = false;
 
             while let Some(event) = stream
@@ -1694,14 +1693,15 @@ impl ApiClient for AnthropicRuntimeClient {
                 match event {
                     ApiStreamEvent::MessageStart(start) => {
                         for block in start.message.content {
-                            push_output_block(block, &mut events, &mut pending_tool, true);
+                            push_output_block(block, 0, &mut events, &mut pending_tools, true);
                         }
                     }
                     ApiStreamEvent::ContentBlockStart(start) => {
                         push_output_block(
                             start.content_block,
+                            start.index,
                             &mut events,
-                            &mut pending_tool,
+                            &mut pending_tools,
                             true,
                         );
                     }
@@ -1712,13 +1712,13 @@ impl ApiClient for AnthropicRuntimeClient {
                             }
                         }
                         ContentBlockDelta::InputJsonDelta { partial_json } => {
-                            if let Some((_, _, input)) = &mut pending_tool {
+                            if let Some((_, _, input)) = pending_tools.get_mut(&delta.index) {
                                 input.push_str(&partial_json);
                             }
                         }
                     },
-                    ApiStreamEvent::ContentBlockStop(_) => {
-                        if let Some((id, name, input)) = pending_tool.take() {
+                    ApiStreamEvent::ContentBlockStop(stop) => {
+                        if let Some((id, name, input)) = pending_tools.remove(&stop.index) {
                             events.push(AssistantEvent::ToolUse { id, name, input });
                         }
                     }
@@ -1839,8 +1839,9 @@ fn convert_messages(messages: &[ConversationMessage]) -> Vec<InputMessage> {
 
 fn push_output_block(
     block: OutputContentBlock,
+    block_index: u32,
     events: &mut Vec<AssistantEvent>,
-    pending_tool: &mut Option<(String, String, String)>,
+    pending_tools: &mut BTreeMap<u32, (String, String, String)>,
     streaming_tool_input: bool,
 ) {
     match block {
@@ -1858,18 +1859,19 @@ fn push_output_block(
             } else {
                 input.to_string()
             };
-            *pending_tool = Some((id, name, initial_input));
+            pending_tools.insert(block_index, (id, name, initial_input));
         }
     }
 }
 
 fn response_to_events(response: MessageResponse) -> Vec<AssistantEvent> {
     let mut events = Vec::new();
-    let mut pending_tool = None;
+    let mut pending_tools = BTreeMap::new();
 
-    for block in response.content {
-        push_output_block(block, &mut events, &mut pending_tool, false);
-        if let Some((id, name, input)) = pending_tool.take() {
+    for (index, block) in response.content.into_iter().enumerate() {
+        let index = u32::try_from(index).expect("response block index overflow");
+        push_output_block(block, index, &mut events, &mut pending_tools, false);
+        if let Some((id, name, input)) = pending_tools.remove(&index) {
             events.push(AssistantEvent::ToolUse { id, name, input });
         }
     }
@@ -2893,6 +2895,7 @@ fn parse_skill_description(contents: &str) -> Option<String> {
 
 #[cfg(test)]
 mod tests {
+    use std::collections::BTreeMap;
     use std::collections::BTreeSet;
     use std::fs;
     use std::io::{Read, Write};
@@ -2905,8 +2908,9 @@ mod tests {
     use super::{
         agent_permission_policy, allowed_tools_for_subagent, execute_agent_with_spawn,
         execute_tool, final_assistant_text, mvp_tool_specs, persist_agent_terminal_state,
-        AgentInput, AgentJob, SubagentToolExecutor,
+        push_output_block, AgentInput, AgentJob, SubagentToolExecutor,
     };
+    use api::OutputContentBlock;
     use runtime::{ApiRequest, AssistantEvent, ConversationRuntime, RuntimeError, Session};
     use serde_json::json;
 
@@ -3121,6 +3125,63 @@ mod tests {
         assert!(error.contains("relative URL without a base") || error.contains("empty host"));
     }
 
+    #[test]
+    fn pending_tools_preserve_multiple_streaming_tool_calls_by_index() {
+        let mut events = Vec::new();
+        let mut pending_tools = BTreeMap::new();
+
+        push_output_block(
+            OutputContentBlock::ToolUse {
+                id: "tool-1".to_string(),
+                name: "read_file".to_string(),
+                input: json!({}),
+            },
+            1,
+            &mut events,
+            &mut pending_tools,
+            true,
+        );
+        push_output_block(
+            OutputContentBlock::ToolUse {
+                id: "tool-2".to_string(),
+                name: "grep_search".to_string(),
+                input: json!({}),
+            },
+            2,
+            &mut events,
+            &mut pending_tools,
+            true,
+        );
+
+        pending_tools
+            .get_mut(&1)
+            .expect("first tool pending")
+            .2
+            .push_str("{\"path\":\"src/main.rs\"}");
+        pending_tools
+            .get_mut(&2)
+            .expect("second tool pending")
+            .2
+            .push_str("{\"pattern\":\"TODO\"}");
+
+        assert_eq!(
+            pending_tools.remove(&1),
+            Some((
+                "tool-1".to_string(),
+                "read_file".to_string(),
+                "{\"path\":\"src/main.rs\"}".to_string(),
+            ))
+        );
+        assert_eq!(
+            pending_tools.remove(&2),
+            Some((
+                "tool-2".to_string(),
+                "grep_search".to_string(),
+                "{\"pattern\":\"TODO\"}".to_string(),
+            ))
+        );
+    }
+
     #[test]
     fn todo_write_persists_and_returns_previous_state() {
         let _guard = env_lock()