feat(llm): support incremental Responses WebSocket transport

.env.example

@@ -9,6 +9,8 @@ OPENAI_TIMEOUT_SECONDS=300
 # 模型请求和模型凭据检查默认直连；设为 true 才读取 Windows/HTTP(S)_PROXY/ALL_PROXY/NO_PROXY。
 # 仅环境配置，不显示在网页中；修改后重启服务生效。TLS 校验及自定义 CA 不受影响。
 OPENAI_USE_SYSTEM_PROXY=false
+# http 保持兼容；websocket 复用连接并按已验证的响应链增量发送，要求网关支持 Responses WebSocket。
+OPENAI_TRANSPORT=http
 # 可选的后台步骤截图摘要单次超时，独立于上面的主模型请求超时。
 ARTEMIS_STEP_SUMMARIZER_TIMEOUT_SECONDS=120
 # 失败后最多重试 5 次，由 Artemis 统一管理，SDK 不叠加重试。

README.md

@@ -61,6 +61,8 @@ cd Artemis-Eapil
 
 也可在项目 `.env` 中设置 `OPENAI_API_KEY`、`OPENAI_BASE_URL` 和 `OPENAI_MODEL`，其他参数见 [.env.example](.env.example)。新安装默认接口地址为 `https://api-n-cd.playones.com/v1`；已保存的地址不变。直接编辑环境文件后需重启相应服务。不要提交自己的密钥。
 
+模型配置的“传输协议”可选择 **WebSocket（增量传输）**，或设置 `OPENAI_TRANSPORT=websocket`。Flash、Pro、探索及辅助模型的异步请求复用 Responses 连接，历史一致时只上传新增输入；网关需支持 `/v1/responses` 的 WebSocket 升级与响应链。默认 HTTP 保持兼容，详细行为与流量诊断见[传输说明](docs/ai-automation.md#responses-websocket-增量传输)。
+
 “功能开关”默认将任务截图长边限制为 2048 像素；关闭后保留设备原始分辨率。可在 [config/artemis.jsonc](config/artemis.jsonc) 的 `agent.screenshot` 中覆盖 `enabled` 和 `max_long_edge`，重启后对新任务生效。
 
 首页“功能配置”提供“Flash 最终 AI 判定与分析”，默认关闭。切换后自动保存并提示结果，失败恢复原值；无需重启服务。开启后，Flash 在执行结束时根据用例需求和执行步骤进行独立复核；关闭后显示“执行完成 · 未做 AI 判定”，保留执行步骤、截图和报告，不自动记为测试通过。此设置适用于所有 Flash 任务，包括批量、精选、MCP/SDK 和探索提炼用例的 Flash 复跑；保存后对尚未开始的任务生效，运行中的任务保持原设置。Pro 和主动探索不受影响。其余功能配置仍需点击底部保存按钮。

apps/admin_console/services/model_settings_service.py

@@ -22,6 +22,7 @@ class ModelSettingsUpdate(BaseModel):
     api_key: SecretStr | None = None  # Omitted/blank keeps the existing key.
     timeout_seconds: float = Field(default=300, gt=0, le=3600)
     max_retries: int = Field(default=5, ge=0, le=10)
+    transport: Literal["http", "websocket"] | None = None
     retry_interval_seconds: float = Field(default=10, ge=0, le=300)
     rate_limit_retry_seconds: float | None = Field(default=None, ge=0, le=3600)
     tool_fallback_model: str | None = Field(default=None, min_length=1, max_length=200)
@@ -72,6 +73,7 @@ def get_model_settings() -> dict:
         "api_key_configured": bool(settings.OPENAI_API_KEY),
         "timeout_seconds": settings.OPENAI_TIMEOUT_SECONDS,
         "max_retries": settings.OPENAI_MAX_RETRIES,
+        "transport": settings.OPENAI_TRANSPORT,
         "retry_interval_seconds": settings.OPENAI_RETRY_INTERVAL_SECONDS,
         "rate_limit_retry_seconds": settings.OPENAI_RATE_LIMIT_RETRY_SECONDS,
         "tool_fallback_model": settings.OPENAI_TOOL_FALLBACK_MODEL,
@@ -102,6 +104,7 @@ def save_model_settings(update: ModelSettingsUpdate) -> dict:
         "OPENAI_REASONING_EFFORT": update.reasoning_effort,
     }
     optional_settings = {
+        "transport": ("OPENAI_TRANSPORT", str),
         "rate_limit_retry_seconds": ("OPENAI_RATE_LIMIT_RETRY_SECONDS", float),
         "tool_fallback_model": ("OPENAI_TOOL_FALLBACK_MODEL", str),
         "tool_fallback_reasoning_effort": ("OPENAI_TOOL_FALLBACK_REASONING_EFFORT", str),
@@ -261,10 +264,19 @@ async def test_model_settings(update: ModelSettingsUpdate) -> dict:
         api_base=update.base_url,
         timeout_seconds=update.timeout_seconds,
         max_retries=0,
+        transport=update.transport or settings.OPENAI_TRANSPORT,
         reasoning_effort=None if update.reasoning_effort == "default" else update.reasoning_effort,
     )
     model = ModelFactory.create_model(endpoint)
+    try:
         result = await model.ainvoke([HumanMessage(content="Reply with OK only.")])
+    finally:
+        await model.http_async_client.aclose()
+        model.http_client.close()
     if not result.content:
         raise ValueError("Responses 接口返回空内容，请检查模型配置")
-    return {"success": True, "message": "Responses 连接测试成功", "model": update.model}
+    return {
+        "success": True,
+        "message": f"Responses {endpoint.transport.upper()} 连接测试成功",
+        "model": update.model,
+    }

apps/showcase_ui/src/app/components/model-settings/model-settings.component.spec.ts

@@ -30,6 +30,31 @@ describe('持久化模型与功能设置', () => {
     http.expectNone('/api/system/model-settings');
   }));
 
+  it('加载、测试和保存 WebSocket 选择，旧配置默认 HTTP', fakeAsync(() => {
+    const fixture = TestBed.createComponent(ModelSettingsComponent);
+    fixture.detectChanges();
+    http.expectOne('/api/system/model-settings').flush(config);
+    tick();
+    fixture.detectChanges();
+    const component = fixture.componentInstance;
+    expect(component.transport).toBe('http');
+    const selector: HTMLSelectElement = fixture.nativeElement.querySelector('#model-transport');
+    selector.value = 'websocket';
+    selector.dispatchEvent(new Event('change'));
+    tick();
+    fixture.detectChanges();
+    component.test();
+    const check = http.expectOne('/api/system/model-settings/test');
+    expect(check.request.body.transport).toBe('websocket');
+    check.flush({message: 'Responses WEBSOCKET 连接测试成功'});
+    component.save();
+    const save = http.expectOne('/api/system/model-settings');
+    expect(save.request.body.transport).toBe('websocket');
+    expect(save.request.body.api_key).toBeUndefined();
+    save.flush({...config, transport: 'websocket'});
+    expect(component.transport).toBe('websocket');
+  }));
+
   it('加载配置后保留密钥，保存新模型且不发送空密钥', () => {
     const fixture = TestBed.createComponent(ModelSettingsComponent);
     fixture.detectChanges();

apps/showcase_ui/src/app/components/model-settings/model-settings.component.ts

@@ -8,6 +8,7 @@ import { ModelSelectorComponent } from '../model-selector/model-selector.compone
 import { MODEL_PRESETS } from '../../core/model-presets';
 
 interface ModelSettings {
+  transport?: 'http' | 'websocket';
   base_url: string;
   model: string;
   api_key_configured: boolean;
@@ -36,6 +37,12 @@ interface ModelSettings {
       <input id="model-base-url" name="baseUrl" type="url" [(ngModel)]="baseUrl" required
         placeholder="https://api-n-cd.playones.com/v1" [disabled]="busy() || loading()" />
       <small>填写接口根地址，通常包含 /v1，无需添加 /responses。</small>
+      <label for="model-transport">传输协议</label>
+      <select id="model-transport" name="transport" [(ngModel)]="transport" [disabled]="busy() || loading()">
+        <option value="http">HTTP（兼容模式）</option>
+        <option value="websocket">WebSocket（增量传输）</option>
+      </select>
+      <small>WebSocket 可减少重复历史与截图的上传，需要接口支持。保存后对新任务生效。</small>
       <app-model-selector label="模型名称" [(value)]="model" [presets]="modelPresets" [disabled]="!!busy() || loading()" />
       <small>选择预设模型，或选择“自定义模型”填写名称。</small>
       <app-setting-slider label="思考强度" [(value)]="reasoning" [values]="reasoningValues" [labels]="reasoningLabels" [disabled]="!!busy() || loading()" hint="更高的强度适合复杂推理；需接口支持，也可选择模型默认。" />
@@ -102,6 +109,7 @@ interface ModelSettings {
   `]
 })
 export class ModelSettingsComponent implements OnInit {
+  transport: 'http' | 'websocket' = 'http';
   readonly reasoningValues = REASONING_VALUES;
   readonly reasoningLabels = REASONING_LABELS;
   private http = inject(HttpClient);
@@ -128,7 +136,7 @@ export class ModelSettingsComponent implements OnInit {
       && Number.isInteger(this.toolFallbackNoProgress) && this.toolFallbackNoProgress >= 1 && this.toolFallbackNoProgress <= 10;
   }
   private payload() {
-    return { base_url: this.baseUrl.trim(), model: this.model.trim(), api_key: this.apiKey.trim() || undefined,
+    return { base_url: this.baseUrl.trim(), transport: this.transport, model: this.model.trim(), api_key: this.apiKey.trim() || undefined,
       timeout_seconds: this.timeout, max_retries: this.retries, retry_interval_seconds: this.retryInterval, reasoning_effort: this.reasoning,
       rate_limit_retry_seconds: this.rateLimitRetry, tool_fallback_model: this.toolFallbackModel.trim(),
       tool_fallback_reasoning_effort: this.toolFallbackReasoning, tool_fallback_max_retries: this.toolFallbackRetries,
@@ -136,6 +144,7 @@ export class ModelSettingsComponent implements OnInit {
   }
   private apply(config: ModelSettings): void {
     this.baseUrl = config.base_url; this.model = config.model; this.timeout = config.timeout_seconds;
+    this.transport = config.transport ?? 'http';
     this.modelPresets = config.model_presets;
     this.retries = config.max_retries; this.retryInterval = config.retry_interval_seconds ?? 10; this.reasoning = config.reasoning_effort ?? 'high'; this.keyConfigured.set(config.api_key_configured); this.envPath.set(config.env_path);
     this.rateLimitRetry = config.rate_limit_retry_seconds ?? 60;

artemis/config/llm.py

@@ -62,6 +62,9 @@ class LLM(BaseModel):
     provider: LLMProvider = "openai"
     model: str = Field(default_factory=lambda: settings.OPENAI_MODEL)
     api_base: str | None = None
+    transport: Literal["http", "websocket"] = Field(
+        default_factory=lambda: settings.OPENAI_TRANSPORT
+    )
     timeout: float = Field(default_factory=lambda: settings.OPENAI_TIMEOUT_SECONDS, gt=0)
     max_retries: int = Field(default_factory=lambda: settings.OPENAI_MAX_RETRIES, ge=0)
     max_tokens: int | None = Field(default=None, gt=0)

artemis/config/settings.py

@@ -16,6 +16,7 @@
 
 import os
 from pathlib import Path
+from typing import Literal
 
 from dotenv import load_dotenv
 from pydantic import Field, SecretStr, model_validator
@@ -109,6 +110,7 @@ class Settings(BaseSettings):
     OPENAI_BASE_URL: str | None = "https://api-n-cd.playones.com/v1"
     OPENAI_MODEL: str = Field(default=DEFAULT_MODEL, min_length=1)
     OPENAI_TIMEOUT_SECONDS: float = Field(default=300.0, gt=0)
+    OPENAI_TRANSPORT: Literal["http", "websocket"] = "http"
     # Environment only; deliberately absent from the saved UI/task configuration.
     OPENAI_USE_SYSTEM_PROXY: bool = False
     OPENAI_MAX_RETRIES: int = Field(default=5, ge=0, le=10)

artemis/llm/diagnostics.py

@@ -390,6 +390,11 @@ async def observe_response_headers(response):
     """Correlate allowlisted headers and passively observe bounded SSE frames."""
     attempt = ACTIVE_ATTEMPT.get()
     if attempt is not None:
+        websocket = response.extensions.get("artemis_websocket", False)
+        if websocket:
+            # 200/SSE is an in-process SDK facade, not a wire HTTP response.
+            attempt.data.update(headers_at=time.time(), phase="waiting_output")
+        else:
             attempt.headers(response)
         content_type = response.headers.get("content-type", "").split(";", 1)[0].lower()
         if (
@@ -407,5 +412,7 @@ async def observe_response_headers(response):
                 )
                 response.stream = stream
                 attempt.wire_streams.append(stream)
+                if websocket:
+                    stream.observation.data["observation_boundary"] = "websocket_event_sse_adapter"
             else:
                 attempt.data["wire_streams_truncated"] = True

artemis/llm/router.py

@@ -17,7 +17,7 @@
 from enum import StrEnum
 import hashlib
 import os
-from typing import Any
+from typing import Any, Literal
 
 from langchain_core.language_models.chat_models import BaseChatModel
 from pydantic import BaseModel, Field
@@ -40,6 +40,9 @@ class ModelEndpoint(BaseModel):
     model_name: str = Field(default_factory=lambda: settings.OPENAI_MODEL)
     api_key: str | None = Field(default=None, repr=False)
     api_base: str | None = None
+    transport: Literal["http", "websocket"] = Field(
+        default_factory=lambda: settings.OPENAI_TRANSPORT
+    )
     temperature: float | None = None
     max_tokens: int | None = Field(default=None, gt=0)
     timeout_seconds: float = Field(default_factory=lambda: settings.OPENAI_TIMEOUT_SECONDS, gt=0)
@@ -63,6 +66,7 @@ class ModelEndpoint(BaseModel):
             self.enable_grounding,
             self.api_base or settings.OPENAI_BASE_URL,
             settings.OPENAI_USE_SYSTEM_PROXY,
+            self.transport,
             hashlib.sha256(key.encode()).hexdigest(),
         )
 
@@ -119,13 +123,32 @@ class ModelFactory:
             "http_client": DefaultHttpxClient(
                 timeout=endpoint.timeout_seconds, **model_http_options()
             ),
-            "http_async_client": DefaultAsyncHttpxClient(
+        }
+        if endpoint.transport == "websocket":
+            import httpx
+
+            from artemis.llm.websocket import ResponsesWebSocketTransport
+
+            # Disable HTTP proxy mounts on this facade: they would bypass the
+            # WebSocket transport. Each actual transport owns its proxy policy.
+            kwargs["http_async_client"] = httpx.AsyncClient(
+                timeout=endpoint.timeout_seconds,
+                trust_env=False,
+                transport=ResponsesWebSocketTransport(
+                    timeout=endpoint.timeout_seconds, use_proxy=settings.OPENAI_USE_SYSTEM_PROXY
+                ),
+                event_hooks={
+                    "request": [observe_request_headers],
+                    "response": [observe_response_headers],
+                },
+            )
+        else:
+            kwargs["http_async_client"] = DefaultAsyncHttpxClient(
                 timeout=endpoint.timeout_seconds,
                 **model_http_options(),
                 event_hooks={
                     "request": [observe_request_headers],
                     "response": [observe_response_headers],
                 },
-            ),
-        }
+            )
         return ChatOpenAI(**{k: v for k, v in kwargs.items() if v is not None})

artemis/llm/websocket.py

@@ -0,0 +1,449 @@
+"""Responses WebSocket transport beneath the existing SDK and safety guards.
+
+The SDK still serializes complete, authoritative requests. Only this wire boundary
+omits a verified prefix. No model messages, tool results, or screenshots are lost
+from local history, so any new connection can start from the full request.
+"""
+
+import asyncio
+from collections import OrderedDict
+from contextlib import suppress
+import hashlib
+import json
+import time
+from uuid import uuid4
+
+import httpx
+from websockets.asyncio.client import connect
+from websockets.exceptions import ConnectionClosed, InvalidStatus, SecurityError
+from websockets.protocol import State
+
+from artemis.llm.diagnostics import ACTIVE_ATTEMPT
+from artemis.llm.http import model_http_options
+from artemis.llm.responses_diagnostics import identifier
+
+IDLE_SECONDS = 60
+MAX_SESSIONS = 32
+MAX_CONNECTION_SECONDS = 55 * 60
+
+
+def _json(value):
+    return json.dumps(value, ensure_ascii=False, separators=(",", ":"), sort_keys=True)
+
+
+def _fingerprint(item):
+    # Assistant message IDs/status are transport metadata; store=false strips
+    # message IDs on replay. Keep every semantic field, including call_id/phase.
+    if isinstance(item, dict):
+        item = dict(item)
+        item.pop("status", None)
+        if item.get("type") == "message" or item.get("role") == "assistant":
+            item.pop("id", None)
+            item.setdefault("type", "message")
+            # Completion events may include token probabilities; LangChain
+            # correctly omits those response-only statistics from replay input.
+            if isinstance(item.get("content"), list):
+                item["content"] = [
+                    {k: v for k, v in block.items() if k != "logprobs"}
+                    if isinstance(block, dict) and block.get("type") == "output_text"
+                    else block
+                    for block in item["content"]
+                ]
+        if item.get("type") == "function_call" and isinstance(item.get("arguments"), str):
+            with suppress(ValueError):
+                item["arguments"] = json.loads(item["arguments"])
+    return hashlib.sha256(_json(item).encode()).digest()
+
+
+class _Connect(connect):
+    def process_redirect(self, exc):
+        # Never forward the API credential to a redirected endpoint.
+        if not isinstance(exc, InvalidStatus) or exc.response.status_code not in {
+            301,
+            302,
+            303,
+            307,
+            308,
+        }:
+            return exc
+        return SecurityError("Responses WebSocket redirects are disabled")
+
+
+class _Session:
+    def __init__(self):
+        self.lock = asyncio.Lock()
+        self.users = 0
+        self.connection = None
+        self.opened_at = 0.0
+        self.previous_id = None
+        self.history = ()
+        self.options = None
+        self.timer = None
+        self.handshake = {}
+
+    async def close(self):
+        if self.timer:
+            self.timer.cancel()
+            self.timer = None
+        self.previous_id = None
+        self.history = ()
+        self.options = None
+        connection, self.connection = self.connection, None
+        if connection is not None:
+            with suppress(ConnectionClosed, OSError, TimeoutError):
+                async with asyncio.timeout(1):
+                    await connection.close()
+
+
+class ResponsesWebSocketTransport(httpx.AsyncBaseTransport):
+    """One serial socket per conversation/role, independent sockets for concurrency."""
+
+    def __init__(self, *, timeout, use_proxy):
+        self.timeout = timeout
+        self.use_proxy = use_proxy
+        self.http = httpx.AsyncClient(timeout=timeout, **model_http_options())
+        self.sessions = OrderedDict()
+        self.cleanup_tasks = set()
+
+    async def aclose(self):
+        sessions = list(self.sessions.values())
+        self.sessions.clear()
+        await asyncio.gather(*(session.close() for session in sessions))
+        if self.cleanup_tasks:
+            await asyncio.gather(*tuple(self.cleanup_tasks), return_exceptions=True)
+        await self.http.aclose()
+
+    def _close_background(self, session):
+        task = asyncio.create_task(session.close())
+        self.cleanup_tasks.add(task)
+        task.add_done_callback(self.cleanup_tasks.discard)
+
+    async def _session(self, request, body):
+        key = body.get("prompt_cache_key")
+        # Header identity prevents credentials, routing, and different gateways
+        # from sharing a socket. Per-request correlation headers aren't affinity.
+        headers = {
+            k: v
+            for k, v in request.headers.items()
+            if k not in {"content-length", "x-client-request-id", "x-client-attempt-id"}
+        }
+        identity = hashlib.sha256(_json([str(request.url), key, headers]).encode()).digest()
+        reusable = isinstance(key, str) and bool(key)
+        if reusable and identity in self.sessions:
+            session = self.sessions.pop(identity)
+            self.sessions[identity] = session
+        else:
+            if len(self.sessions) >= MAX_SESSIONS:
+                for old_key, old in list(self.sessions.items()):
+                    if not old.users:
+                        del self.sessions[old_key]
+                        # Reserve the replacement before yielding: concurrent
+                        # requests for its scope must find the same session.
+                        self._close_background(old)
+                        break
+            session = _Session()
+            reusable = reusable and len(self.sessions) < MAX_SESSIONS
+            if reusable:
+                self.sessions[identity] = session
+        acquired = False
+        session.users += 1
+        try:
+            await session.lock.acquire()
+            acquired = True
+        finally:
+            if not acquired:
+                session.users -= 1
+                self._idle(identity, session, reusable)
+        if session.timer:
+            session.timer.cancel()
+            session.timer = None
+        return identity, session, reusable
+
+    def release(self, identity, session, reusable):
+        session.lock.release()
+        session.users -= 1
+        self._idle(identity, session, reusable)
+
+    def _idle(self, identity, session, reusable):
+        if not reusable or session.users or self.sessions.get(identity) is not session:
+            return
+
+        def expire():
+            if session.users or self.sessions.get(identity) is not session:
+                return
+            del self.sessions[identity]
+            self._close_background(session)
+
+        session.timer = asyncio.get_running_loop().call_later(IDLE_SECONDS, expire)
+
+    async def handle_async_request(self, request):
+        if request.method != "POST" or not request.url.path.rstrip("/").endswith("/responses"):
+            return await self.http.send(request, stream=True)
+        body = json.loads(await request.aread())
+        identity, session, reusable = await self._session(request, body)
+        exchange = _Exchange(self, request, body, identity, session, reusable)
+        stream_handed_off = False
+        try:
+            await exchange.start()
+            first = await exchange.read()
+            if first.get("type") == "error":
+                status = first.get("status", 400)
+                status = status if type(status) is int and 400 <= status <= 599 else 400
+                return httpx.Response(status, json={"error": first.get("error", {})})
+            if not body.get("stream"):
+                event = first
+                while not exchange.done:
+                    event = await exchange.read()
+                    if event.get("type") == "error":
+                        raise httpx.ReadError("Responses WebSocket request failed", request=request)
+                return httpx.Response(
+                    200, json=event["response"], extensions={"artemis_websocket": True}
+                )
+            response = httpx.Response(
+                200,
+                headers={"content-type": "text/event-stream"},
+                stream=_EventStream(exchange, first),
+                extensions={"artemis_websocket": True},
+            )
+            stream_handed_off = True
+            return response
+        except InvalidStatus as error:
+            return httpx.Response(
+                error.response.status_code,
+                json={
+                    "error": {
+                        "message": "Responses WebSocket 升级失败，请检查网关 WebSocket 支持或选择 HTTP。"
+                    }
+                },
+                headers={
+                    k: v
+                    for k, v in error.response.headers.raw_items()
+                    if k.lower() in {"retry-after", "x-request-id"}
+                },
+            )
+        finally:
+            if not stream_handed_off:
+                await exchange.aclose()
+
+
+class _Exchange:
+    def __init__(self, transport, request, body, identity, session, reusable):
+        self.transport = transport
+        self.request = request
+        self.body = body
+        self.identity = identity
+        self.session = session
+        self.reusable = reusable
+        self.done = False
+        self.closed = False
+        self.recovered = False
+        self.output_seen = False
+        self.input = body.get("input")
+        self.input_hashes = (
+            tuple(map(_fingerprint, self.input)) if isinstance(self.input, list) else ()
+        )
+        self.options = hashlib.sha256(
+            _json(
+                {
+                    k: v
+                    for k, v in body.items()
+                    if k not in {"input", "stream", "background", "previous_response_id"}
+                }
+            ).encode()
+        ).digest()
+        self.explicit_state = bool(body.get("previous_response_id") or body.get("conversation"))
+        self.attempt = ACTIVE_ATTEMPT.get()
+        self.stats = {
+            "mode": "websocket",
+            "full_request_bytes": len(request.content),
+            "sent_payload_bytes": 0,
+            "received_payload_bytes": 0,
+            "create_events": 0,
+            "incremental": False,
+            "reused_input_items": 0,
+            "connection_reused": session.connection is not None,
+        }
+        if self.attempt is not None:
+            self.attempt.data["transport"] = self.stats
+
+    async def start(self):
+        session = self.session
+        if session.connection is not None and (
+            session.connection.state is not State.OPEN
+            or time.monotonic() - session.opened_at >= MAX_CONNECTION_SECONDS
+        ):
+            await session.close()
+            self.stats["connection_reused"] = False
+        if session.connection is None:
+            url = self.request.url.copy_with(
+                scheme="wss" if self.request.url.scheme == "https" else "ws"
+            )
+            headers = {
+                k: v
+                for k, v in self.request.headers.items()
+                if k
+                not in {"host", "content-length", "content-type", "connection", "accept-encoding"}
+            }
+            options = {"ssl": model_http_options()["verify"]} if url.scheme == "wss" else {}
+            session.connection = await _Connect(
+                str(url),
+                additional_headers=headers,
+                user_agent_header=None,
+                proxy=True if self.transport.use_proxy else None,
+                open_timeout=self.transport.timeout,
+                close_timeout=1,
+                max_size=64 * 1024 * 1024,
+                **options,
+            )
+            session.opened_at = time.monotonic()
+            session.handshake = {
+                "client_request_id": identifier(headers.get("x-client-request-id")),
+                "client_attempt_id": identifier(headers.get("x-client-attempt-id")),
+            }
+        if self.attempt:
+            self.attempt.data.update(session.handshake, request_header_scope="websocket_handshake")
+            for field, local in (
+                ("client_request_id", "request_id"),
+                ("client_attempt_id", "attempt_id"),
+            ):
+                self.attempt.data[field + "_matches"] = session.handshake.get(
+                    field
+                ) == self.attempt.data.get(local)
+        self.stats["handshake_http_status"] = 101
+        await self.send(incremental=True)
+
+    async def send(self, *, incremental):
+        session = self.session
+        payload = {k: v for k, v in self.body.items() if k not in {"stream", "background"}}
+        count = len(session.history)
+        can_continue = (
+            incremental
+            and not self.explicit_state
+            and session.previous_id
+            and session.options == self.options
+            and count
+            and len(self.input_hashes) > count
+            and self.input_hashes[:count] == session.history
+        )
+        if can_continue:
+            payload["previous_response_id"] = session.previous_id
+            payload["input"] = self.input[count:]
+            self.stats.update(incremental=True, reused_input_items=count)
+        else:
+            self.stats.update(incremental=False, reused_input_items=0)
+            self.stats["full_context_reason"] = (
+                "explicit_state"
+                if self.explicit_state
+                else "history_or_options_changed"
+                if session.previous_id
+                else "new_connection_or_chain"
+            )
+        payload["type"] = "response.create"
+        encoded = _json(payload)
+        await session.connection.send(encoded)
+        self.stats["sent_payload_bytes"] += len(encoded.encode())
+        self.stats["create_events"] += 1
+        self.stats["payload_bytes_saved"] = max(
+            0, self.stats["full_request_bytes"] - self.stats["sent_payload_bytes"]
+        )
+        if self.attempt:
+            self.attempt.data["request_sent_at"] = time.time()
+            self.attempt.publish_if_due()
+
+    async def read(self):
+        try:
+            async with asyncio.timeout(self.transport.timeout):
+                raw = await self.session.connection.recv()
+            self.stats["received_payload_bytes"] += len(
+                raw.encode() if isinstance(raw, str) else raw
+            )
+            event = json.loads(raw)
+            if not isinstance(event, dict):
+                raise ValueError("Responses WebSocket event is not an object")
+        except ConnectionClosed as error:
+            raise httpx.ReadError(
+                "Responses WebSocket connection closed before completion", request=self.request
+            ) from error
+        except (ValueError, UnicodeError) as error:
+            raise httpx.ReadError(
+                "Invalid Responses WebSocket event", request=self.request
+            ) from error
+        kind = event.get("type", "")
+        error = event.get("error") or {}
+        if (
+            kind == "error"
+            and error.get("code") == "previous_response_not_found"
+            and self.stats["incremental"]
+            and not self.recovered
+            and not self.output_seen
+        ):
+            # An explicit rejection means no model output was generated. Recover
+            # once from the authoritative input; never replay an ambiguous drop.
+            self.recovered = True
+            self.session.previous_id = None
+            self.stats["chain_recovered"] = True
+            await self.send(incremental=False)
+            return await self.read()
+        if kind not in {"error", "response.created", "response.in_progress", "response.queued"}:
+            self.output_seen = True
+        if kind in {
+            "response.completed",
+            "response.failed",
+            "response.incomplete",
+            "response.cancelled",
+        }:
+            self.done = True
+            response = event.get("response") or {}
+            self.session.previous_id = None
+            if (
+                kind == "response.completed"
+                and response.get("status") == "completed"
+                and isinstance(response.get("id"), str)
+                and response["id"]
+                and isinstance(response.get("output"), list)
+                and self.input_hashes
+                and not self.explicit_state
+            ):
+                output = [
+                    item
+                    for item in response["output"]
+                    if not (
+                        item.get("type") == "reasoning"
+                        and not item.get("encrypted_content")
+                        and self.body.get("store") is False
+                    )
+                ]
+                self.session.previous_id = response["id"]
+                self.session.history = self.input_hashes + tuple(map(_fingerprint, output))
+                self.session.options = self.options
+        return event
+
+    async def aclose(self):
+        if self.closed:
+            return
+        self.closed = True
+        try:
+            if not self.done or not self.reusable:
+                await self.session.close()
+        finally:
+            self.transport.release(self.identity, self.session, self.reusable)
+
+
+class _EventStream(httpx.AsyncByteStream):
+    def __init__(self, exchange, first):
+        self.exchange = exchange
+        self.first = first
+
+    async def __aiter__(self):
+        try:
+            event = self.first
+            while True:
+                yield b"data: " + _json(event).encode() + b"\n\n"
+                if self.exchange.done or event.get("type") == "error":
+                    return
+                event = await self.exchange.read()
+        finally:
+            await self.aclose()
+
+    async def aclose(self):
+        await self.exchange.aclose()

artemis/resources/config/artemis.jsonc

@@ -2,6 +2,7 @@
 {
   "default": {
     "provider": "openai",
+    "transport": "${env:OPENAI_TRANSPORT:-http}",
     "model": "${env:OPENAI_MODEL:-gpt-6-luna}",
     "timeout": "${env:OPENAI_TIMEOUT_SECONDS:-300}",
     "max_retries": "${env:OPENAI_MAX_RETRIES:-5}",

artemis/services/llm.py

@@ -1174,6 +1174,7 @@ def _resolve_endpoint(
         else settings.OPENAI_MAX_RETRIES,
         max_tokens=_get_val(cfg, "max_tokens", int),
         api_base=_get_val(cfg, "api_base", str),
+        transport=_get_val(cfg, "transport", str) or settings.OPENAI_TRANSPORT,
         reasoning_effort=_get_val(cfg, "reasoning_effort", str),
         enable_grounding=_get_val(cfg, "enable_grounding", bool) or False,
     )

config/artemis.jsonc

@@ -2,6 +2,7 @@
 {
   "default": {
     "provider": "openai",
+    "transport": "${env:OPENAI_TRANSPORT:-http}",
     "model": "${env:OPENAI_MODEL:-gpt-6-luna}",
     "timeout": "${env:OPENAI_TIMEOUT_SECONDS:-300}",
     "max_retries": "${env:OPENAI_MAX_RETRIES:-5}",

docs/ai-automation.md

@@ -169,6 +169,18 @@ Flash、Pro（启用 transcript）和探索共用分阶段上下文。一个阶
 
 辅助摘要继续独立处理各自片段，避免为了提高命中比例累积不相关的历史。视觉摘要中完全重复的插件执行结果与控件树只发送一份；独立错误、状态、不同观察及原始 trace 均保留。
 
+## Responses WebSocket 增量传输
+
+网页“模型配置 → 传输协议”选择 `WebSocket（增量传输）`，或在 `.env` 设置 `OPENAI_TRANSPORT=websocket`。网页保存对新任务生效；直接编辑 `.env` 后重启相应服务。默认 `http`，旧配置不变。JSON 模型/节点配置可用 `transport` 单独覆盖；MCP 使用 `mobile_config(section="model", changes={"transport":"websocket"}, persist_defaults=true)` 保存，仍须用户明确要求修改默认值。
+
+异步任务请求在现有 SDK 序列化与完整性保护之下接入 `/v1/responses` WebSocket。首轮发送完整上下文；同一任务（或探索）、角色、模型、路由头及凭据范围内复用连接。只在当前输入严格包含上一轮输入和完整输出、生成选项一致时，发送 `previous_response_id` 与新增输入。仅忽略消息 ID、状态、输出概率等不影响历史语义的响应字段，不按文本相似度截取。工具结果保留 `call_id`，文本、图片、结构化输出与 usage 继续走原有解析和安全检查。同步底层模型调用仍使用 HTTP；Flash、Pro、探索和辅助模型运行路径均为异步。
+
+本地历史、截图替换或压缩后重新发送当前完整上下文；不会为了续传保留已经移除的旧界面索引。连接重建或服务端明确返回 `previous_response_not_found` 时，从本地完整历史恢复，后者最多补发一次且必须尚未出现输出。连接在输出中途断开时，不在传输层盲目重发，交给现有模型重试预算；取消关闭未完成连接，不采纳半截工具参数或结论。不同会话并行、同一连接串行；连接池最多保留 32 个会话，空闲 60 秒释放，连接满 55 分钟主动重建。`store=false` 保持不变。
+
+网关必须支持 WebSocket Upgrade、双向长连接、`response.create` 与响应 ID 续接；只透传会话头不能证明支持。握手传递现有授权和路由头，保持 TLS/自定义 CA 验证与 `OPENAI_USE_SYSTEM_PROXY` 策略，拒绝携带凭据跨地址重定向。不支持升级时明确报错，可在网页改回 HTTP，不会静默改成 HTTP 全量请求。
+
+请求诊断的 `transport` 记录 `full_request_bytes`、`sent_payload_bytes`、`received_payload_bytes`、`payload_bytes_saved`、`incremental`、`reused_input_items`、`connection_reused`、`create_events` 和重建原因。字节数衡量 JSON 应用载荷，不包含 TLS/WebSocket 帧和压缩开销；SDK 内部兼容 SSE 的字节不冒充网络流量或 HTTP 200。握手上的请求标识属于建连尝试，后续轮次以本地请求诊断及响应 ID 关联。减少上传不等于减少模型上下文 token，也不等于缓存命中，usage 仍以服务端实报为准。协议见 [OpenAI 官方说明](https://developers.openai.com/api/docs/guides/websocket-mode)。
+
 ## 指定任务诊断、复盘与管理
 
 流式请求中的 `click`、`long_press`、`swipe`、`press_key`、`manage_app`、`wait_for_delay` 每次调用参数超过 16,384 字符时，会立即关闭该请求并丢弃未完成输出，避免持续生成异常参数直到超时。该保护只针对小型设备动作，不限制输入文本、笔记、报告或插件参数，也不改变模型思考强度与输出 token 配置。异常最多额外尝试一次（配置为不重试时不重试）；仍异常时进入现有备用模型或暂停恢复流程。诊断记录保留 `invalid_output` 分类及 `stream_guard` 中的工具名、实际字符数和保护上限，不保存参数正文；无法据此断定异常来自模型生成还是 API 中转。

mcp_server/catalog.json

@@ -437,7 +437,7 @@
     },
     {
       "name": "mobile_config",
-      "description": "Read configuration or explicitly save non-secret defaults for FUTURE tasks.\n\nOnly set persist_defaults=true when the user requests saved/global configuration.\nmodel changes: model (preset or custom), base_url, reasoning_effort\n(default/none/minimal/low/medium/high/xhigh/max), timeout_seconds, max_retries,\nretry_interval_seconds, rate_limit_retry_seconds, tool_fallback_model,\ntool_fallback_reasoning_effort, tool_fallback_max_retries,\ntool_fallback_no_progress_attempts. 'default' omits reasoning;\nunsupported values fail visibly.\ncapabilities: video_enabled and recording_enabled are independent; outputter_enabled,\nflash_final_check_enabled (false; Flash exit-only AI review),\nstep_summarizer_enabled, screenshot_limit_enabled, task_timeout_seconds (default 1800),\nobservation_timeout_seconds (60), observation_max_attempts (3),\ndevice_command_timeout_seconds (10), action_failure_attempts (3),\nworker_cleanup_timeout_seconds (30), adb_auto_connect (true),\nadb_discovery_interval_seconds (10), adb_discovery_timeout_seconds (5).\nreports: formats=['docx','xlsx'] (at least one).\nexploration: model (default gpt-6.1-terra), reasoning_effort (high), timeout_seconds\n(1800 active seconds), demonstration_timeout_seconds (900), demonstration_warning_seconds (60).\nAPI keys are never accepted or returned: use the Web UI or .env for credentials.\nSaving settings does not change an already running task. Flash/Pro are execution\nprofiles, not model names. Model retries do not repeat device test cases.\n",
+      "description": "Read configuration or explicitly save non-secret defaults for FUTURE tasks.\n\nOnly set persist_defaults=true when the user requests saved/global configuration.\nmodel changes: model (preset or custom), base_url, transport (http/websocket), reasoning_effort\n(default/none/minimal/low/medium/high/xhigh/max), timeout_seconds, max_retries,\nretry_interval_seconds, rate_limit_retry_seconds, tool_fallback_model,\ntool_fallback_reasoning_effort, tool_fallback_max_retries,\ntool_fallback_no_progress_attempts. 'default' omits reasoning;\nunsupported values fail visibly.\ncapabilities: video_enabled and recording_enabled are independent; outputter_enabled,\nflash_final_check_enabled (false; Flash exit-only AI review),\nstep_summarizer_enabled, screenshot_limit_enabled, task_timeout_seconds (default 1800),\nobservation_timeout_seconds (60), observation_max_attempts (3),\ndevice_command_timeout_seconds (10), action_failure_attempts (3),\nworker_cleanup_timeout_seconds (30), adb_auto_connect (true),\nadb_discovery_interval_seconds (10), adb_discovery_timeout_seconds (5).\nreports: formats=['docx','xlsx'] (at least one).\nexploration: model (default gpt-6.1-terra), reasoning_effort (high), timeout_seconds\n(1800 active seconds), demonstration_timeout_seconds (900), demonstration_warning_seconds (60).\nAPI keys are never accepted or returned: use the Web UI or .env for credentials.\nSaving settings does not change an already running task. Flash/Pro are execution\nprofiles, not model names. Model retries do not repeat device test cases.\n",
       "inputSchema": {
         "properties": {
           "section": {

mcp_server/tools/automation.py

@@ -14,6 +14,7 @@ from mcp_server.base import mcp
 
 EXPORT_DIR = ROOT_DIR / "artifacts" / "mcp_exports"
 MODEL_FIELDS = {
+    "transport",
     "base_url",
     "model",
     "timeout_seconds",
@@ -186,7 +187,7 @@ async def mobile_config(
     """Read configuration or explicitly save non-secret defaults for FUTURE tasks.
 
     Only set persist_defaults=true when the user requests saved/global configuration.
-    model changes: model (preset or custom), base_url, reasoning_effort
+    model changes: model (preset or custom), base_url, transport (http/websocket), reasoning_effort
     (default/none/minimal/low/medium/high/xhigh/max), timeout_seconds, max_retries,
     retry_interval_seconds, rate_limit_retry_seconds, tool_fallback_model,
     tool_fallback_reasoning_effort, tool_fallback_max_retries,

pyproject.toml

@@ -57,7 +57,7 @@ dependencies = [
     "uvicorn[standard]>=0.30.1",
     "httpx>=0.28.1",
     "requests>=2.31.0",
-    "websockets>=12.0",
+    "websockets>=15.0",
     "sseclient-py>=1.8.0",
     # Image, Video & Multimodal Perception
     "pillow>=13.0.0",

skills/artemis-use/references/modes.md

@@ -16,6 +16,8 @@
 
 ## 模式、模型与配置
 
+网页模型配置和 `mobile_config(section="model")` 提供 `transport=http/websocket`。用户明确要求保存默认值时可设置 `changes={"transport":"websocket"}, persist_defaults=true`；新任务的 Flash、Pro、探索和辅助请求共用此选择。WebSocket 在历史一致时仅上传新增输入，网关须支持 Responses WebSocket；压缩、重连会发送完整上下文，上传减少不能当作 token 减少或缓存命中。不支持协议时应查看错误并确认网关能力，不自动改写用户配置。
+
 - `mobile_run_task.model` 取 `Flash` / `Pro`。API 模型名放 `runtime_overrides.model`；思考强度放 `runtime_overrides.reasoning_effort`。批次使用 `task_config.profile` 选择模式，`task_config.model` 选择 API 模型。
 - 探索独立配置，默认 `gpt-6.1-terra/high`，支持 `gpt-6-astra` 及已配置网关支持的自定义模型。执行、示范分析、提炼共用该探索配置；不改变 Flash/Pro 默认值。
 - 默认 high、单次模型请求时限 300 秒。`default` 表示不发送思考强度，`none` 是显式关闭，不能混为一谈；接口不支持参数时如实报告，不擅自降级。

tests/unit/admin_console/test_model_settings.py

@@ -21,6 +21,7 @@ def isolated_settings(monkeypatch, tmp_path):
         "OPENAI_MODEL",
         "OPENAI_BASE_URL",
         "OPENAI_TIMEOUT_SECONDS",
+        "OPENAI_TRANSPORT",
         "OPENAI_MAX_RETRIES",
         "OPENAI_RETRY_INTERVAL_SECONDS",
         "OPENAI_RATE_LIMIT_RETRY_SECONDS",
@@ -50,6 +51,7 @@ def isolated_settings(monkeypatch, tmp_path):
 
 def test_save_reload_and_blank_key_preserves_secret(isolated_settings):
     update = service.ModelSettingsUpdate(
+        transport="websocket",
         base_url="https://gateway.example/v1/",
         model="my-model",
         api_key="new-test-secret",
@@ -63,6 +65,9 @@ def test_save_reload_and_blank_key_preserves_secret(isolated_settings):
     )
     result = service.save_model_settings(update)
     assert result["api_key_configured"]
+    assert result["transport"] == "websocket"
+    assert dotenv_values(isolated_settings)["OPENAI_TRANSPORT"] == "websocket"
+    assert service.ModelSettingsUpdate(model="test").transport is None
     assert "new-test-secret" not in repr(result)
     saved = dotenv_values(isolated_settings)
     assert saved["OPENAI_API_KEY"] == "new-test-secret"

tests/unit/mcp/test_automation_tools.py

@@ -72,6 +72,8 @@ async def test_config_requires_explicit_persistence_and_excludes_secrets(monkeyp
     payload = request.call_args.kwargs["json"]
     assert payload["rate_limit_retry_seconds"] == 90
     assert payload["tool_fallback_model"] == "gpt-6-luna"
+    await tools.mobile_config(changes={"transport": "websocket"}, persist_defaults=True)
+    assert request.call_args.kwargs["json"]["transport"] == "websocket"
 
 
 @pytest.mark.asyncio

tests/unit/test_responses_websocket.py

@@ -0,0 +1,605 @@
+"""Real local WebSocket/SDK round trips; no device, credential, or paid endpoint."""
+
+import asyncio
+from contextlib import asynccontextmanager
+from copy import deepcopy
+import json
+
+import httpx
+from langchain_core.messages import HumanMessage, ToolMessage
+import pytest
+import pytest_asyncio
+from websockets.asyncio.server import serve
+
+from artemis.llm.diagnostics import ACTIVE_ATTEMPT
+from artemis.llm.response_guard import ResponseIntegrityError, ensure_response_complete
+from artemis.llm.router import ModelEndpoint, ModelFactory
+from artemis.services.llm import RobustChatModelWrapper
+
+
+def text_output(text="OK"):
+    return [
+        {
+            "id": "msg_test",
+            "type": "message",
+            "role": "assistant",
+            "status": "completed",
+            "content": [{"type": "output_text", "text": text, "annotations": [], "logprobs": []}],
+        }
+    ]
+
+
+class Wire:
+    def __init__(self):
+        self.requests = []
+        self.connections = []
+        self.outputs = []
+        self.reject_previous = False
+        self.drop = False
+        self.stall = False
+        self.status = "completed"
+        self.sent = asyncio.Event()
+        self.gate = None
+
+    async def handle(self, socket):
+        self.connections.append(socket)
+        async for raw in socket:
+            body = json.loads(raw)
+            self.requests.append(body)
+            self.sent.set()
+            if self.reject_previous and body.get("previous_response_id"):
+                await socket.send(
+                    json.dumps(
+                        {
+                            "type": "error",
+                            "status": 400,
+                            "error": {"code": "previous_response_not_found"},
+                        }
+                    )
+                )
+                continue
+            response = {
+                "id": f"resp_{len(self.requests)}",
+                "object": "response",
+                "created_at": 1,
+                "status": "in_progress",
+                "model": "test-model",
+                "output": [],
+            }
+            await socket.send(
+                json.dumps({"type": "response.created", "sequence_number": 0, "response": response})
+            )
+            if self.drop:
+                await socket.close()
+                continue
+            if self.stall:
+                await socket.wait_closed()
+                return
+            if self.gate:
+                await self.gate.wait()
+            output = self.outputs.pop(0) if self.outputs else text_output()
+            seq = 0
+            for index, item in enumerate(output):
+                seq += 1
+                await socket.send(
+                    json.dumps(
+                        {
+                            "type": "response.output_item.added",
+                            "sequence_number": seq,
+                            "output_index": index,
+                            "item": {**item, "content": [], "arguments": ""},
+                        }
+                    )
+                )
+                for content_index, block in enumerate(item.get("content", [])):
+                    if block.get("type") != "output_text":
+                        continue
+                    seq += 1
+                    await socket.send(
+                        json.dumps(
+                            {
+                                "type": "response.output_text.delta",
+                                "sequence_number": seq,
+                                "output_index": index,
+                                "content_index": content_index,
+                                "item_id": item["id"],
+                                "delta": block["text"],
+                            }
+                        )
+                    )
+                    seq += 1
+                    await socket.send(
+                        json.dumps(
+                            {
+                                "type": "response.output_text.done",
+                                "sequence_number": seq,
+                                "output_index": index,
+                                "content_index": content_index,
+                                "item_id": item["id"],
+                                "text": block["text"],
+                            }
+                        )
+                    )
+                seq += 1
+                await socket.send(
+                    json.dumps(
+                        {
+                            "type": "response.output_item.done",
+                            "sequence_number": seq,
+                            "output_index": index,
+                            "item": item,
+                        }
+                    )
+                )
+            response.update(
+                status=self.status,
+                output=output,
+                usage={
+                    "input_tokens": 20,
+                    "output_tokens": 3,
+                    "total_tokens": 23,
+                    "input_tokens_details": {"cached_tokens": 10},
+                },
+            )
+            await socket.send(
+                json.dumps(
+                    {
+                        "type": f"response.{self.status}",
+                        "sequence_number": seq + 1,
+                        "response": response,
+                    }
+                )
+            )
+
+
+@pytest_asyncio.fixture
+async def wire(monkeypatch):
+    monkeypatch.delenv("ARTEMIS_FAKE_LLM", raising=False)
+    instance = Wire()
+    async with serve(instance.handle, "127.0.0.1", 0) as server:
+        instance.base = f"http://127.0.0.1:{server.sockets[0].getsockname()[1]}/v1"
+        yield instance
+
+
+@asynccontextmanager
+async def model(wire, **kwargs):
+    endpoint = ModelEndpoint(
+        model_name="test-model",
+        api_key="test-secret",
+        api_base=wire.base,
+        transport="websocket",
+        max_retries=0,
+        timeout_seconds=2,
+        **kwargs,
+    )
+    value = ModelFactory.create_model(endpoint)
+    try:
+        yield value
+    finally:
+        await value.http_async_client.aclose()
+        value.http_client.close()
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("stream", [False, True])
+async def test_sdk_roundtrip_incremental_history_and_usage(wire, stream):
+    async with model(wire) as raw:
+        messages = [
+            HumanMessage(
+                content=[
+                    {"type": "text", "text": "Inspect"},
+                    {
+                        "type": "image_url",
+                        "image_url": {"url": "data:image/png;base64," + "A" * 80000},
+                    },
+                ]
+            )
+        ]
+
+        async def call(history):
+            if not stream:
+                return await raw.ainvoke(history, prompt_cache_key="task:operator")
+            chunks = [
+                chunk async for chunk in raw.astream(history, prompt_cache_key="task:operator")
+            ]
+            return sum(chunks[1:], chunks[0])
+
+        first = await call(messages)
+        ensure_response_complete(first)
+        second = await call([*messages, first, HumanMessage(content="Continue")])
+        ensure_response_complete(second)
+        assert second.usage_metadata["input_token_details"]["cache_read"] == 10
+        assert wire.requests[1].get("previous_response_id") == "resp_1", {
+            "first": first.content,
+            "tail": wire.requests[1]["input"][1:],
+        }
+        assert wire.requests[1]["input"] == [
+            {"type": "message", "role": "user", "content": "Continue"}
+        ]
+        assert "stream" not in wire.requests[0]
+        assert wire.requests[0]["type"] == "response.create"
+        assert wire.requests[0]["store"] is False
+        assert len(json.dumps(wire.requests[1])) < len(json.dumps(wire.requests[0])) / 100
+        assert len(wire.connections) == 1
+        assert wire.connections[0].request.path == "/v1/responses"
+        assert wire.connections[0].request.headers["Authorization"] == "Bearer test-secret"
+
+
+@pytest.mark.asyncio
+async def test_tool_call_chain_preserves_call_id(wire):
+    wire.outputs.append(
+        [
+            {
+                "id": "fc_test",
+                "type": "function_call",
+                "call_id": "call_test",
+                "name": "lookup",
+                "arguments": '{"x":1}',
+                "status": "completed",
+            }
+        ]
+    )
+    async with model(wire) as raw:
+        messages = [HumanMessage(content="Look up")]
+        first = await raw.ainvoke(messages, prompt_cache_key="task:operator")
+        await raw.ainvoke(
+            [*messages, first, ToolMessage(content="found", tool_call_id="call_test")],
+            prompt_cache_key="task:operator",
+        )
+        assert wire.requests[1]["previous_response_id"] == "resp_1"
+        assert wire.requests[1]["input"] == [
+            {"type": "function_call_output", "call_id": "call_test", "output": "found"}
+        ]
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    "change", ["history", "options", "scope", "new_connection", "server_closed"]
+)
+async def test_chain_resets_on_context_or_connection_change(wire, change):
+    async with model(wire) as raw:
+        messages = [HumanMessage(content="Original")]
+        first = await raw.ainvoke(messages, prompt_cache_key="task:operator")
+        history = [*messages, first, HumanMessage(content="Next")]
+        options = {"prompt_cache_key": "task:operator"}
+        if change == "history":
+            history[0] = HumanMessage(content="Compacted")
+        if change == "options":
+            options["temperature"] = 0.3
+        if change == "scope":
+            options["prompt_cache_key"] = "task:checker"
+        if change == "new_connection":
+            await raw.http_async_client._transport.sessions[
+                next(iter(raw.http_async_client._transport.sessions))
+            ].close()
+        if change == "server_closed":
+            await wire.connections[0].close()
+            await wire.connections[0].wait_closed()
+        await raw.ainvoke(history, **options)
+        assert "previous_response_id" not in wire.requests[1]
+        assert len(wire.requests[1]["input"]) == 3
+
+
+@pytest.mark.asyncio
+async def test_uncached_id_recovers_once_with_full_context(wire):
+    async with model(wire) as raw:
+        history = [HumanMessage(content="First")]
+        first = await raw.ainvoke(history, prompt_cache_key="task:operator")
+        wire.reject_previous = True
+        result = await raw.ainvoke(
+            [*history, first, HumanMessage(content="Next")], prompt_cache_key="task:operator"
+        )
+        ensure_response_complete(result)
+        assert len(wire.requests) == 3
+        assert wire.requests[1]["previous_response_id"] == "resp_1"
+        assert "previous_response_id" not in wire.requests[2]
+        assert len(wire.requests[2]["input"]) == 3
+
+
+@pytest.mark.asyncio
+async def test_missing_affinity_never_shares_connections(wire):
+    async with model(wire) as raw:
+        history = [HumanMessage(content="First")]
+        first = await raw.ainvoke(history)
+        await raw.ainvoke([*history, first, HumanMessage(content="Next")])
+        assert len(wire.connections) == 2
+        assert "previous_response_id" not in wire.requests[1]
+
+
+@pytest.mark.asyncio
+async def test_cancel_closes_socket_and_next_call_starts_full(wire):
+    async with model(wire) as raw:
+        wire.stall = True
+        task = asyncio.create_task(
+            raw.ainvoke([HumanMessage(content="Wait")], prompt_cache_key="task:operator")
+        )
+        await asyncio.wait_for(wire.sent.wait(), 2)
+        task.cancel()
+        with pytest.raises(asyncio.CancelledError):
+            await task
+        await asyncio.wait_for(wire.connections[0].wait_closed(), 2)
+        wire.stall = False
+        await raw.ainvoke([HumanMessage(content="Next")], prompt_cache_key="task:operator")
+        assert len(wire.connections) == 2
+        assert "previous_response_id" not in wire.requests[1]
+
+
+@pytest.mark.asyncio
+async def test_disconnect_never_silently_resends_model_request(wire):
+    async with model(wire) as raw:
+        wire.drop = True
+        with pytest.raises(Exception, match="connection|Connection"):
+            await raw.ainvoke([HumanMessage(content="Test")], prompt_cache_key="task:operator")
+        assert len(wire.requests) == 1
+
+
+@pytest.mark.asyncio
+async def test_incomplete_result_does_not_advance_chain(wire):
+    async with model(wire) as raw:
+        wire.status = "incomplete"
+        first = await raw.ainvoke([HumanMessage(content="First")], prompt_cache_key="task:operator")
+        with pytest.raises(ResponseIntegrityError):
+            ensure_response_complete(first)
+        wire.status = "completed"
+        await raw.ainvoke(
+            [HumanMessage(content="First"), first, HumanMessage(content="Next")],
+            prompt_cache_key="task:operator",
+        )
+        assert "previous_response_id" not in wire.requests[1]
+
+
+@pytest.mark.asyncio
+async def test_diagnostics_measure_actual_wire_payload_without_logging_content(wire):
+    class Attempt:
+        data = {}
+
+        def publish_if_due(self):
+            pass
+
+    attempt = Attempt()
+    async with model(wire) as raw:
+        # Disable the unrelated HTTP trace hook: this checks the transport's
+        # own byte counters, alongside the real SDK request serialization.
+        raw.http_async_client.event_hooks = {"request": [], "response": []}
+        first = await raw.ainvoke(
+            [HumanMessage(content="A" * 5000)], prompt_cache_key="task:operator"
+        )
+        token = ACTIVE_ATTEMPT.set(attempt)
+        try:
+            await raw.ainvoke(
+                [HumanMessage(content="A" * 5000), first, HumanMessage(content="Next")],
+                prompt_cache_key="task:operator",
+            )
+        finally:
+            ACTIVE_ATTEMPT.reset(token)
+        stats = deepcopy(attempt.data["transport"])
+        assert stats["incremental"] is True
+        assert stats["sent_payload_bytes"] == len(
+            json.dumps(
+                wire.requests[1], ensure_ascii=False, separators=(",", ":"), sort_keys=True
+            ).encode()
+        )
+        assert stats["payload_bytes_saved"] > 4000
+        assert "AAAA" not in json.dumps(stats)
+
+
+@pytest.mark.asyncio
+async def test_gateway_rejects_upgrade_with_explicit_error(wire):
+    import openai
+    from websockets.http11 import Response
+    from websockets.datastructures import Headers
+
+    async with serve(
+        wire.handle,
+        "127.0.0.1",
+        0,
+        process_request=lambda *_: Response(404, "Not Found", Headers(), b""),
+    ) as server:
+        wire.base = f"http://127.0.0.1:{server.sockets[0].getsockname()[1]}/v1"
+        async with model(wire) as raw:
+            with pytest.raises(openai.NotFoundError, match="WebSocket"):
+                await raw.ainvoke([HumanMessage(content="Test")])
+
+
+@pytest.mark.asyncio
+async def test_wrapper_retains_stream_guards_and_diagnostic_events(wire, monkeypatch):
+    from artemis.services import llm
+    from tests.unit.test_responses_wire_lifecycle import Engine
+
+    engine = Engine()
+    monkeypatch.setattr(llm, "_get_current_data_engine", lambda: engine)
+    async with model(wire) as raw:
+        wrapper = RobustChatModelWrapper(
+            raw,
+            endpoint=ModelEndpoint(model_name="test-model", transport="websocket", max_retries=0),
+        )
+        result = await wrapper.ainvoke(
+            [HumanMessage(content="Test")], prompt_cache_key="task:operator"
+        )
+        ensure_response_complete(result)
+        assert engine.data["transport"]["mode"] == "websocket"
+        assert engine.data["transport"]["received_payload_bytes"] > 0
+        assert engine.data.get("http_status") is None
+
+
+@pytest.mark.asyncio
+async def test_parallel_scopes_keep_independent_response_chains(wire):
+    async with model(wire) as raw:
+
+        async def conversation(scope):
+            history = [HumanMessage(content=f"First {scope}")]
+            first = await raw.ainvoke(history, prompt_cache_key=scope)
+            await raw.ainvoke(
+                [*history, first, HumanMessage(content=f"Next {scope}")], prompt_cache_key=scope
+            )
+            return scope, first.response_metadata["id"]
+
+        results = await asyncio.gather(
+            conversation("task-a:operator"), conversation("task-b:operator")
+        )
+        assert len(wire.connections) == 2
+        for scope, response_id in results:
+            calls = [item for item in wire.requests if item["prompt_cache_key"] == scope]
+            assert len(calls) == 2
+            assert calls[1]["previous_response_id"] == response_id
+            assert len(calls[1]["input"]) == 1
+
+
+@pytest.mark.asyncio
+async def test_idle_connections_are_closed_and_removed(wire, monkeypatch):
+    from artemis.llm import websocket
+
+    monkeypatch.setattr(websocket, "IDLE_SECONDS", 0.01)
+    async with model(wire) as raw:
+        await raw.ainvoke([HumanMessage(content="Test")], prompt_cache_key="task:operator")
+        await asyncio.wait_for(wire.connections[0].wait_closed(), 2)
+        assert not raw.http_async_client._transport.sessions
+
+
+@pytest.mark.asyncio
+async def test_cancel_waiting_turn_keeps_active_chain_and_idle_cleanup(wire, monkeypatch):
+    from artemis.llm import websocket
+
+    monkeypatch.setattr(websocket, "IDLE_SECONDS", 0.05)
+    wire.gate = asyncio.Event()
+    async with model(wire) as raw:
+        history = [HumanMessage(content="First")]
+        first = asyncio.create_task(raw.ainvoke(history, prompt_cache_key="task:operator"))
+        await asyncio.wait_for(wire.sent.wait(), 2)
+        queued = asyncio.create_task(raw.ainvoke(history, prompt_cache_key="task:operator"))
+        async with asyncio.timeout(2):
+            while next(iter(raw.http_async_client._transport.sessions.values())).users != 2:
+                await asyncio.sleep(0)
+        queued.cancel()
+        with pytest.raises(asyncio.CancelledError):
+            await queued
+        wire.gate.set()
+        result = await first
+        await raw.ainvoke(
+            [*history, result, HumanMessage(content="Next")], prompt_cache_key="task:operator"
+        )
+        assert len(wire.connections) == 1
+        assert len(wire.requests) == 2
+        assert wire.requests[1]["previous_response_id"] == result.response_metadata["id"]
+        await asyncio.wait_for(wire.connections[0].wait_closed(), 2)
+        assert not raw.http_async_client._transport.sessions
+
+
+@pytest.mark.asyncio
+async def test_pool_eviction_does_not_split_concurrent_scope(wire, monkeypatch):
+    from artemis.llm import websocket
+
+    monkeypatch.setattr(websocket, "MAX_SESSIONS", 1)
+    closing, resume = asyncio.Event(), asyncio.Event()
+    async with model(wire) as raw:
+        await raw.ainvoke([HumanMessage(content="Old")], prompt_cache_key="old")
+        old = next(iter(raw.http_async_client._transport.sessions.values()))
+        close = old.close
+
+        async def delayed_close():
+            closing.set()
+            await resume.wait()
+            await close()
+
+        monkeypatch.setattr(old, "close", delayed_close)
+        first = asyncio.create_task(
+            raw.ainvoke([HumanMessage(content="New first")], prompt_cache_key="new")
+        )
+        try:
+            await asyncio.wait_for(closing.wait(), 2)
+            await asyncio.wait_for(
+                asyncio.gather(
+                    first,
+                    raw.ainvoke([HumanMessage(content="New second")], prompt_cache_key="new"),
+                ),
+                2,
+            )
+            assert len(wire.connections) == 2
+            assert len(wire.requests) == 3
+        finally:
+            resume.set()
+
+
+@pytest.mark.asyncio
+async def test_structured_output_uses_same_websocket_path(wire):
+    from pydantic import BaseModel
+
+    class Result(BaseModel):
+        ok: bool
+
+    wire.outputs.append(
+        [
+            {
+                "id": "fc_test",
+                "type": "function_call",
+                "call_id": "call_test",
+                "name": "Result",
+                "arguments": '{"ok":true}',
+                "status": "completed",
+            }
+        ]
+    )
+    async with model(wire) as raw:
+        result = (
+            await RobustChatModelWrapper(
+                raw, endpoint=ModelEndpoint(transport="websocket", max_retries=0)
+            )
+            .with_structured_output(Result)
+            .ainvoke([HumanMessage(content="Test")])
+        )
+        assert result.ok is True
+        assert len(wire.requests) == 1
+
+
+@pytest.mark.asyncio
+async def test_invalid_tool_arguments_are_rejected_before_use(wire, monkeypatch):
+    from artemis.services import llm
+    from artemis.llm.reliability import LLMExhaustedError
+
+    monkeypatch.setattr(llm, "_get_current_data_engine", lambda: None)
+    wire.outputs.append(
+        [
+            {
+                "id": "fc_test",
+                "type": "function_call",
+                "call_id": "call_test",
+                "name": "lookup",
+                "arguments": '{"x":',
+                "status": "completed",
+            }
+        ]
+    )
+    async with model(wire) as raw:
+        wrapper = RobustChatModelWrapper(
+            raw, endpoint=ModelEndpoint(transport="websocket", max_retries=0)
+        )
+        with pytest.raises(LLMExhaustedError):
+            await wrapper.ainvoke([HumanMessage(content="Test")], prompt_cache_key="task:operator")
+        assert len(wire.requests) == 1
+
+
+@pytest.mark.asyncio
+async def test_redirect_does_not_forward_authorization(wire):
+    from websockets.http11 import Response
+    from websockets.datastructures import Headers
+
+    async with serve(
+        wire.handle,
+        "127.0.0.1",
+        0,
+        process_request=lambda *_: Response(
+            307,
+            "Redirect",
+            Headers({"Location": wire.base.replace("http://", "ws://") + "/responses"}),
+            b"",
+        ),
+    ) as server:
+        redirect_base = f"http://127.0.0.1:{server.sockets[0].getsockname()[1]}/v1"
+        original_base, wire.base = wire.base, redirect_base
+        try:
+            async with model(wire) as raw:
+                with pytest.raises(Exception, match="redirects are disabled"):
+                    await raw.ainvoke([HumanMessage(content="Test")])
+            assert not wire.connections
+        finally:
+            wire.base = original_base

uv.lock

@@ -369,7 +369,7 @@ requires-dist = [
     { name = "uiautomator2", specifier = ">=3.5.0" },
     { name = "uuid-utils", specifier = ">=0.9.0" },
     { name = "uvicorn", extras = ["standard"], specifier = ">=0.30.1" },
-    { name = "websockets", specifier = ">=12.0" },
+    { name = "websockets", specifier = ">=15.0" },
     { name = "windows-capture", marker = "platform_machine == 'AMD64' and sys_platform == 'win32'", specifier = ">=2.0.1,<3" },
 ]
 provides-extras = ["dev"]