@linxiraos/pi-ai 1.1.4 → 1.1.6

package/CHANGELOG.md

@@ -2,11 +2,13 @@
 
 ## [Unreleased]
 
-## [1.1.4] - 2026-08-26
+## [1.1.6] - 2026-08-30
 
-### Changed
+- 同步上游 OMP v18.0.9（`cc14e04f075d`）。
+
+## [1.1.5] - 2026-08-26
 
-- 同步 1.1.4 发布线（与 1.1.3 无功能差异）。
+- 同步上游 OMP v18.0.5 / v18.0.6：新增 Yolo-Auto / OpenRouter 浏览器登录与 DeepInfra 支持，空补全重试重构（withReplaySafeStreamRetry）。
 
 ## [1.1.3] - 2026-08-25
 

package/README.md

@@ -60,6 +60,7 @@ Unified LLM API with automatic model discovery, provider configuration, token an
 - **NVIDIA** (requires `NVIDIA_API_KEY`)
 - **NanoGPT** (requires `NANO_GPT_API_KEY`)
 - **Novita** (requires `NOVITA_API_KEY`)
+- **DeepInfra** (requires `DEEPINFRA_API_KEY`)
 - **Hugging Face Inference**
 - **xAI**
 - **Venice** (requires `VENICE_API_KEY`)
@@ -75,7 +76,7 @@ Unified LLM API with automatic model discovery, provider configuration, token an
 - **Qwen Portal** (supports `QWEN_OAUTH_TOKEN` or `QWEN_PORTAL_API_KEY`)
 - **QwenCloud Token Plan** (supports `/login alibaba-token-plan`, `ALIBABA_TOKEN_PLAN_API_KEY`, or `BAILIAN_TOKEN_PLAN_API_KEY`; interactive login first selects a region — International (Singapore, default), China (Beijing) for 百炼 Token Plan keys, or a custom base URL — since region keys are non-interchangeable, then optionally stores a `home.qwencloud.com` Cookie request header for best-effort 5-hour and 7-day quota reporting)
   To enable quota reporting, sign in to the Token Plan dashboard, copy the `Cookie` request-header value from a `home.qwencloud.com` request in browser developer tools, and paste it at the second login prompt. Press Enter to skip; the Cookie is sensitive and session-lived, so rerun login when it expires.
-- **Cloudflare AI Gateway** (requires `CLOUDFLARE_AI_GATEWAY_API_KEY` and provider-specific gateway base URL)
+- **Cloudflare AI Gateway** (supports `/login cloudflare-ai-gateway`, or `CLOUDFLARE_AI_GATEWAY_API_KEY` with `CLOUDFLARE_ACCOUNT_ID` and `CLOUDFLARE_GATEWAY_ID`)
 - **Ollama** (local OpenAI-compatible runtime; optional `OLLAMA_API_KEY`)
 - **Ollama Cloud** (hosted native Ollama API; requires `OLLAMA_CLOUD_API_KEY`)
 - **llama.cpp** (local OpenAI and Anthropic compatible inference server)
@@ -943,6 +944,7 @@ In Node.js environments, you can set environment variables to avoid passing API
 | NVIDIA         | `NVIDIA_API_KEY`                                                             |
 | NanoGPT        | `NANO_GPT_API_KEY`                                                          |
 | Novita         | `NOVITA_API_KEY`                                                           |
+| DeepInfra      | `DEEPINFRA_API_KEY`                                                          |
 | Venice         | `VENICE_API_KEY`                                                             |
 | Moonshot       | `MOONSHOT_API_KEY`                                                           |
 | xAI            | `XAI_API_KEY`                                                                |
@@ -958,11 +960,10 @@ In Node.js environments, you can set environment variables to avoid passing API
 | Xiaomi MiMo    | `XIAOMI_API_KEY`                                                             |
 | ZenMux         | `ZENMUX_API_KEY`                                                             |
 | vLLM           | `VLLM_API_KEY`                                                               |
-| Cloudflare AI Gateway | `CLOUDFLARE_AI_GATEWAY_API_KEY`                                      |
+| Cloudflare AI Gateway | `CLOUDFLARE_AI_GATEWAY_API_KEY` + `CLOUDFLARE_ACCOUNT_ID` + `CLOUDFLARE_GATEWAY_ID` |
 | GitHub Copilot | `COPILOT_GITHUB_TOKEN` or `GH_TOKEN` or `GITHUB_TOKEN`                      |
 
-For Cloudflare AI Gateway models, use provider base URL format
-`https://gateway.ai.cloudflare.com/v1/<account>/<gateway>/anthropic`.
+`/login cloudflare-ai-gateway` collects and stores the gateway token, account ID, and gateway ID. For environment configuration, set all three Cloudflare values above. OMP derives provider endpoints from the account and gateway IDs.
 
 For Anthropic Foundry routing, set `CLAUDE_CODE_USE_FOUNDRY=true` plus:
 `FOUNDRY_BASE_URL`, `ANTHROPIC_FOUNDRY_API_KEY`, optional `ANTHROPIC_CUSTOM_HEADERS`,
@@ -983,6 +984,7 @@ Provider endpoint defaults for the current OpenAI-compatible integrations:
 - NVIDIA: `https://integrate.api.nvidia.com/v1`
 - NanoGPT: `https://nano-gpt.com/api/v1`
 - Novita: `https://api.novita.ai/openai/v1`
+- DeepInfra: `https://api.deepinfra.com/v1/openai`
 - Hugging Face Inference: `https://router.huggingface.co/v1`
 - Venice: `https://api.venice.ai/api/v1`
 - Xiaomi MiMo: `https://api.xiaomimimo.com/anthropic`
@@ -993,7 +995,7 @@ Provider endpoint defaults for the current OpenAI-compatible integrations:
 - Ollama: local OpenAI-compatible runtime (`http://127.0.0.1:11434/v1`)
 - Ollama Cloud: native Ollama API host (`https://ollama.com/api`, configured here as base URL `https://ollama.com`)
 - LiteLLM: `http://localhost:4000/v1`
-- Cloudflare AI Gateway: `https://gateway.ai.cloudflare.com/v1/<account>/<gateway>/anthropic`
+- Cloudflare AI Gateway: native Anthropic, OpenAI, and Workers AI routes under `https://gateway.ai.cloudflare.com/v1/<account>/<gateway>`
 - Qwen Portal: `https://portal.qwen.ai/v1`
 When set, the library automatically uses these keys:
 
@@ -1084,7 +1086,7 @@ Credentials are saved to `agent.db` in the agent directory. `/login qianfan` ope
 
 `login` supports OAuth providers (Anthropic, OpenAI Codex, GitHub Copilot, Gemini CLI, Antigravity) and API-key onboarding flows.
 
-For the current API-key onboarding flows, the library covers Together, Moonshot, Qianfan, NVIDIA, NanoGPT, Novita, Hugging Face, Venice, Xiaomi, vLLM, LiteLLM, Cloudflare AI Gateway, Qwen Portal, and Ollama Cloud. Ollama remains the local runtime integration; set `OLLAMA_API_KEY` only when your local or self-hosted deployment enforces bearer auth.
+For the current API-key onboarding flows, the library covers Together, Moonshot, Qianfan, NVIDIA, NanoGPT, Novita, DeepInfra, Hugging Face, Venice, Xiaomi, vLLM, LiteLLM, Cloudflare AI Gateway, Qwen Portal, and Ollama Cloud. Ollama remains the local runtime integration; set `OLLAMA_API_KEY` only when your local or self-hosted deployment enforces bearer auth.
 
 ### Programmatic OAuth
 

package/dist/types/auth-broker/remote-store.d.ts

@@ -1,7 +1,7 @@
 import { type AuthCredential, type AuthCredentialStore, type DisabledCredentialSummary, type OAuthCredential, type StoredAuthCredential, type StoredCredentialBlock } from "../auth-storage.js";
 import type { OAuthCredentials } from "../registry/oauth/types.js";
 import type { Provider } from "../types.js";
-import type { ObservedUsageEntry, UsageReport } from "../usage.js";
+import type { ClientUsageIdentity, ObservedUsageEntry, UsageReport } from "../usage.js";
 import { type AuthBrokerClient } from "./client.js";
 import type { SnapshotResponse } from "./types.js";
 /**
@@ -35,6 +35,14 @@ export interface RemoteAuthCredentialStoreOptions {
     accountPool?: AuthBrokerAccountPool;
     /** Flush cadence for batched observed-usage reports. Default 10s. */
     observedUsageFlushMs?: number;
+    /**
+     * Idle window after the last foreground store use before background
+     * snapshot sync (SSE stream / long-poll) disconnects and parks. A parked
+     * store holds no timers or sockets, so an unclosed store never keeps the
+     * process alive longer than one idle window. Sync resumes transparently on
+     * the next use. Default 20s.
+     */
+    backgroundIdleMs?: number;
 }
 export declare class RemoteAuthCredentialStore implements AuthCredentialStore {
     #private;
@@ -130,7 +138,10 @@ export declare class RemoteAuthCredentialStore implements AuthCredentialStore {
      * a flush. One `POST /v1/usage/observed` at most per flush interval; on
      * failure the batch is retained and retried with the next flush. A 404
      * (pre-endpoint broker) disables reporting for the life of this store.
+     *
+     * `client` overrides the reporting identity — the auth-gateway attributes
+     * each request to the originating install/app instead of the gateway host.
      */
-    recordObservedUsage(entries: ObservedUsageEntry[]): void;
+    recordObservedUsage(entries: ObservedUsageEntry[], client?: ClientUsageIdentity): void;
     close(): void;
 }

package/dist/types/auth-gateway/http.d.ts

@@ -1,4 +1,5 @@
 import type { Api, AssistantMessage, Model } from "../types.js";
+import type { ClientUsageIdentity } from "../usage.js";
 export declare function json(status: number, body: unknown, headers?: Record<string, string>): Response;
 /**
  * Diagnostic response headers for translated inference requests, mirroring the
@@ -30,6 +31,19 @@ export declare function isAuthorized(req: Request, tokens: ReadonlySet<string>):
  * `handleFormatEndpoint`; parsers then read `options.headers`.
  */
 export declare function captureRequestHeaders(headers: Headers): Record<string, string>;
+/**
+ * Resolve the usage-attribution identity for an inbound gateway request.
+ *
+ * pi-native omp clients send `x-omp-install-id` / `x-omp-hostname` /
+ * `x-omp-app` (see `providers/pi-native-client.ts`); any client may set them.
+ * Requests without an install id fall back to the gateway host's identity
+ * under the `gateway` app label, so unlabeled foreign-SDK traffic (llm-git,
+ * openai/anthropic SDKs) still lands in per-client burn tracking instead of
+ * vanishing. These headers are attribution-only — they are deliberately
+ * absent from {@link captureRequestHeaders}'s allow-list and never reach the
+ * upstream provider.
+ */
+export declare function resolveClientIdentity(headers: Headers): ClientUsageIdentity;
 /**
  * Resolve a prompt-cache identity from inbound request body + headers.
  * Order of precedence (first wins):

package/dist/types/auth-storage.d.ts

@@ -1,7 +1,7 @@
 import type { ApiKeyResolver } from "./auth-retry.js";
 import type { OAuthAuthInfo, OAuthController, OAuthCredentials, OAuthProviderId } from "./registry/oauth/types.js";
 import type { Provider } from "./types.js";
-import type { ClientUsageReport, ClientUsageSummary, CredentialRankingStrategy, ObservedUsageEntry, UsageHistoryEntry, UsageHistoryQuery, UsageLogger, UsageProvider, UsageReport } from "./usage.js";
+import type { ClientUsageIdentity, ClientUsageReport, ClientUsageSummary, CredentialRankingStrategy, ObservedUsageEntry, UsageHistoryEntry, UsageHistoryQuery, UsageLogger, UsageProvider, UsageReport } from "./usage.js";
 import { type CodexResetConsumeCode, type CodexResetCredit } from "./usage/openai-codex-reset.js";
 export { isSqliteBusyError, isSqliteCorruptionError, SqliteAuthCredentialStore, } from "./auth/sqlite-credential-store.js";
 export type ApiKeyCredential = {
@@ -317,8 +317,10 @@ export interface AuthCredentialStore {
      * Client hook: forward locally observed request usage. Remote broker stores
      * batch these to the broker so it can attribute token burn per install;
      * local stores omit it and observation is skipped.
+     * `client` overrides the reporting identity (gateway requests attribute to
+     * the originating client, not the gateway host).
      */
-    recordObservedUsage?(entries: ObservedUsageEntry[]): void;
+    recordObservedUsage?(entries: ObservedUsageEntry[], client?: ClientUsageIdentity): void;
     /** Broker host: persist one client's observed-usage report. */
     recordClientUsage?(report: ClientUsageReport): void;
     /** Broker host: aggregate recorded per-client usage since a timestamp. */
@@ -819,6 +821,18 @@ export declare class AuthStorage {
      * `XAI_API_KEY` does not auto-select SuperGrok (`xai-oauth`).
      */
     hasAuth(provider: string): boolean;
+    /**
+     * Like {@link hasAuth} but excludes providers whose only credential is the
+     * self-resolving {@link AUTHENTICATED_SENTINEL} — the marker AWS/Vertex
+     * transports return when a credential *source* merely exists (a stray
+     * `~/.aws` profile, an EC2 instance role, Application Default Credentials)
+     * without a usable key resolved yet. Default-model auto-selection uses this
+     * so an ambiently-available provider (e.g. `amazon-bedrock` via an unrelated
+     * AWS profile) does not win the startup default over a provider the user
+     * actually signed into and then 403 on the first turn. Explicit selection
+     * and picker visibility still go through {@link hasAuth}. See issue #9967.
+     */
+    hasConcreteAuth(provider: string): boolean;
     /**
      * Whether a request could resolve a key for this provider, including
      * cross-provider env aliases (`xai-oauth` borrowing `XAI_API_KEY`).
@@ -915,6 +929,8 @@ export declare class AuthStorage {
         };
         costUsd?: number;
         at?: number;
+        /** Attribution override; defaults to this process's install identity. */
+        client?: ClientUsageIdentity;
     }): void;
     /** Broker host: persist one client's observed-usage report (per-install token burn). */
     recordClientUsage(report: ClientUsageReport): boolean;

package/dist/types/providers/anthropic.d.ts

@@ -218,11 +218,8 @@ export declare function isInvalidThinkingSignatureError(message: string): boolea
  */
 export declare function maybeAddReplayUnsignedThinkingHint(model: Model<"anthropic-messages">, message: string): string;
 /**
- * Public entry: wrap the single-attempt streamer with bounded empty-completion
- * retries (a benign terminal stop carrying no content/usage would otherwise
- * stall the agent loop). The inner attempt keeps its own provider-failure retry
- * loop; this layer only re-issues a fresh request on an empty success. Shared
- * with the OpenAI-completions provider via `withEmptyCompletionRetry`.
+ * Public entry: retry benign empty completions before they reach the agent
+ * loop. The inner attempt owns Anthropic provider-failure retries.
  */
 export declare const streamAnthropic: StreamFunction<"anthropic-messages">;
 export type AnthropicSystemBlock = {

package/dist/types/providers/azure-openai-responses.d.ts

@@ -12,6 +12,9 @@ export interface AzureOpenAIResponsesOptions extends StreamOptions {
     disableReasoning?: boolean;
 }
 /**
- * Generate function for Azure OpenAI Responses API
+ * Retries transient Azure stream failures only before assistant output commits
+ * the attempt. The unsupported explicit prompt-cache config is rejected
+ * synchronously here — callers of the direct entrypoint get the immediate
+ * `ConfigurationError` rather than a stream whose `.result()` rejects later.
  */
 export declare const streamAzureOpenAIResponses: StreamFunction<"azure-openai-responses">;

package/dist/types/providers/claude-code-fingerprint.d.ts

@@ -8,9 +8,11 @@
  * provider module.
  */
 /** Claude runtime version bundled by the current Cowork desktop release. */
-export declare const claudeCodeVersion = "2.1.220";
+export declare const claudeCodeVersion = "2.1.246";
+/** `@anthropic-ai/sdk` version bundled by the current Cowork desktop release. */
+export declare const claudeCodeSdkVersion = "0.112.1";
 /** User-Agent emitted by Cowork's `claude-desktop` inference entrypoint. */
-export declare const coworkUserAgent = "claude-cli/2.1.220 (external, claude-desktop)";
+export declare const coworkUserAgent = "claude-cli/2.1.246 (external, claude-desktop)";
 /** Prefix used to isolate custom Anthropic OAuth tools from built-in tools. */
 export declare const claudeToolPrefix: string;
 /** Identity block prepended by Cowork's Claude runtime. */

package/dist/types/providers/openai-completions.d.ts

@@ -38,10 +38,8 @@ export interface OpenAICompletionsOptions extends StreamOptions {
     promptCache?: OpenAIPromptCacheOptions;
 }
 /**
- * Public entry: wrap the single-attempt streamer with bounded empty-completion
- * retries — flaky gateways occasionally 200 with `delta: {}` + `finish_reason:
- * "stop"` and no usage, which would otherwise stall the agent loop. Shared with
- * the Anthropic provider via `withEmptyCompletionRetry`.
+ * Retries benign empty completions and transient provider failures only before
+ * assistant output commits the attempt.
  */
 export declare const streamOpenAICompletions: StreamFunction<"openai-completions">;
 export declare function parseChunkUsage(rawUsage: object, model: Model<"openai-completions">, premiumRequests: number | undefined): AssistantMessage["usage"];

package/dist/types/providers/openai-data-uri.d.ts

@@ -0,0 +1,11 @@
+/** Base64 payload and media type decoded from a data URI. */
+export interface DecodedDataUri {
+    data: string;
+    mimeType: string;
+}
+/**
+ * Decodes base64 and percent-encoded `data:` URIs.
+ *
+ * Returns `undefined` for non-data URLs and data URIs without a comma separator.
+ */
+export declare function decodeDataUri(url: string): DecodedDataUri | undefined;