@anuma/sdk 6.39.0 → 6.39.2

package/dist/expo/index.d.ts

@@ -2002,6 +2002,127 @@ declare function createStreamingDeAnonymizer(redactor: PiiRedactor, emit: (chunk
     flush: () => void;
 };
 
+type LlmTokenUsage = {
+    inputTokens?: number;
+    outputTokens?: number;
+};
+type ModelCallStartEvent = {
+    runId: string;
+    stepIndex: number;
+    model: string;
+    messages: Array<LlmapiMessage>;
+    tools: Array<LlmapiChatCompletionTool>;
+    requestBody: Record<string, unknown>;
+};
+type ModelCallEndEvent = {
+    runId: string;
+    stepIndex: number;
+    content: string;
+    toolCalls: Array<{
+        id: string;
+        name: string;
+        arguments: string;
+    }>;
+    usage?: LlmTokenUsage;
+    finishReason?: string;
+    error?: string;
+};
+type ToolUseStartEvent = {
+    runId: string;
+    stepIndex: number;
+    toolCallId: string;
+    name: string;
+    rawArguments: string;
+    parsedArguments?: Record<string, unknown>;
+};
+type ToolUseEndEvent = {
+    runId: string;
+    stepIndex: number;
+    toolCallId: string;
+    name: string;
+    result?: unknown;
+    error?: string;
+    errorType?: "parse" | "timeout" | "execution" | "cancelled";
+};
+type RunStartEvent = {
+    runId: string;
+    model: string;
+    messages: Array<LlmapiMessage>;
+    tools: Array<LlmapiChatCompletionTool>;
+};
+type RunEndEvent = {
+    runId: string;
+    finalContent: string;
+    /** Number of LLM rounds executed. */
+    totalSteps: number;
+};
+type RunErrorEvent = {
+    runId: string;
+    error: string;
+    /**
+     * Only `"model"` today. Tool failures surface on {@link RunHooks.afterToolUse}
+     * with `errorType`, not on `onRunError`, so a tool stage isn't needed yet.
+     * Left as a union to leave room for `"validation"` or `"tool"` later
+     * without a breaking change.
+     */
+    stage: "model";
+    /**
+     * The underlying `Error` when one is available. `RunErrorEvent` predates
+     * this field, so `error` (a string message) is kept for backwards
+     * compatibility; new consumers should prefer `errorObject` when present
+     * for stack traces and `cause`.
+     */
+    errorObject?: Error;
+};
+type RunHooks = {
+    /**
+     * Fires once per `runToolLoop` invocation, before any pre-processor or
+     * model call. The `messages` payload is the raw caller-supplied
+     * conversation — pre-processors that enrich it (e.g. web search context)
+     * run afterwards, so `beforeModelCall.messages` reflects the
+     * post-enrichment state.
+     */
+    onRunStart?: (e: RunStartEvent) => Promise<void> | void;
+    /**
+     * Fires exactly once per run on clean completion. Mutually exclusive
+     * with `onRunError` — a run that completes fires `onRunEnd`; a run that
+     * errors or is aborted fires `onRunError`. Consumers can rely on
+     * exactly one terminal hook per `onRunStart`.
+     */
+    onRunEnd?: (e: RunEndEvent) => Promise<void> | void;
+    /**
+     * Fires exactly once per run when the loop errors or is aborted.
+     * `stage` is `"model"` for stream/provider errors and for aborts;
+     * tool execution failures surface as structured `{ error, errorType }`
+     * results on `afterToolUse` rather than firing `onRunError`.
+     * Mutually exclusive with `onRunEnd`.
+     */
+    onRunError?: (e: RunErrorEvent) => Promise<void> | void;
+    beforeModelCall?: (e: ModelCallStartEvent) => Promise<void> | void;
+    afterModelCall?: (e: ModelCallEndEvent) => Promise<void> | void;
+    /**
+     * Fires before each tool the model invoked — including server-side
+     * tools that have no client executor. Today this is observe-only.
+     * TODO(follow-up PR): allow returning { args?, abort?: { reason } } to
+     * mutate arguments or short-circuit the tool call.
+     *
+     * Pairing with {@link afterToolUse} is asymmetric: server-side tools
+     * routed via `onToolCall` get `beforeToolUse` but no `afterToolUse`,
+     * because they don't execute in-process. Tools blocked by a failed
+     * dependency or a dependency cycle DO get `afterToolUse` with
+     * `errorType: "execution"` — consumers tallying matched pairs should
+     * filter on executor-backed tools to avoid spurious imbalances.
+     */
+    beforeToolUse?: (e: ToolUseStartEvent) => Promise<void> | void;
+    /**
+     * Fires after each executor-backed tool finishes (success or failure)
+     * and after each dependency-blocked tool is skipped (`errorType:
+     * "execution"`). Does NOT fire for server-side tools routed via
+     * `onToolCall`. See {@link beforeToolUse} for the pairing contract.
+     */
+    afterToolUse?: (e: ToolUseEndEvent) => Promise<void> | void;
+};
+
 /**
  * Configuration for stream output smoothing.
  *
@@ -2196,6 +2317,8 @@ type BaseSendMessageArgs = ResponsesApiOptions & {
 type BaseUseChatOptions = {
     getToken?: () => Promise<string | null>;
     baseUrl?: string;
+    /** Run lifecycle hooks for telemetry and tracing, for example from `createMetricsHooks` */
+    hooks?: RunHooks | RunHooks[];
     /**
      * Callback function to be called when a new data chunk is received.
      */
@@ -6615,6 +6738,8 @@ interface CachedServerTools {
     version: string;
     /** Checksum from the server for cache invalidation */
     checksum?: string;
+    /** ETag from the server. The next refresh sends it in an If-None-Match header. */
+    etag?: string;
 }
 /**
  * Pluggable persistence for the fetched server-tools list.
@@ -6684,7 +6809,9 @@ declare function clearServerToolsCache(cache: ToolsCacheBackend): void | Promise
  * Flow:
  * 1. Check the cache backend (localStorage by default; override via `cache`)
  * 2. If cache valid and not force refresh, return cached tools
- * 3. Otherwise, fetch from API, cache, and return
+ * 3. Otherwise, fetch from API, cache, and return. When the cache holds an ETag,
+ *    the request sends If-None-Match. On 304 the cached tools stay and only the
+ *    stored timestamp changes.
  * 4. On fetch failure, return cached tools if available (stale-while-error)
  */
 declare function getServerTools(options: ServerToolsOptions): Promise<ServerTool[]>;
@@ -7131,6 +7258,8 @@ interface CreateConversationOptions {
  * @inline
  */
 interface BaseUseChatStorageOptions {
+    /** Run lifecycle hooks for telemetry and tracing, for example from `createMetricsHooks` */
+    hooks?: RunHooks | RunHooks[];
     /** WatermelonDB database instance for storing conversations and messages */
     database: Database;
     /** Route memory reads, writes, recall and the memory tools through this store instead of the database's vault; folders are unsupported. With `piiRedaction` on, build the store with `embeddingOptions.maskInput` (e.g. `(text) => redactor.maskText(text)` with any `PiiRedactor`) or its memory calls throw. */
@@ -10551,6 +10680,8 @@ type UseCreditsResult = {
  */
 declare function useCredits(options?: UseCreditsOptions): UseCreditsResult;
 
+/** Clear all cached model lists. */
+declare function clearModelsCache(): void;
 /**
  * @inline
  */
@@ -11147,4 +11278,4 @@ interface UseChatStorageResult extends BaseUseChatStorageResult {
  */
 declare function useChatStorage(options: UseChatStorageOptions): UseChatStorageResult;
 
-export { type AutoExtractMessage, type AutoExtractor, type Budget, CHUNKS_DISCARDED_ORIGIN, type CachedChunkVectors, type CachedServerTools, Conversation as ChatConversation, Message as ChatMessage, type ChatRole, type ChunkVectorCache, type ChunkingOptions, type ConsolidationAction, type ConsolidationFallbackReason, type ConversationMemoryInput, ConversationMemory as ConversationMemoryModel, type ConversationMemoryOperationsContext, type CreateAutoExtractorOptions, type CreateConversationOptions, type CreateDecaySweeperOptions, type CreateMessageOptions, type CreateVaultFolderOptions, type CreateVaultMemoryOptions, type CryptoPriceClassification, type CryptoPricePreProcessorOptions, DEFAULT_CACHE_EXPIRATION_MS, DEFAULT_CHUNK_CACHE_SIZE, DEFAULT_CHUNK_OVERLAP, DEFAULT_CHUNK_SIZE, DEFAULT_DECAY_POLICY, DEFAULT_MAX_CLASSIFIER_CALLS_PER_SWEEP, DEFAULT_MIN_CHUNK_SIZE, DEFAULT_PROFILE_FACETS, DEFAULT_PROFILE_FACT_TYPE_WEIGHTS, DEFAULT_PROFILE_PROOF_ALPHA, DEFAULT_PROFILE_TREND_MULTIPLIERS, DEFAULT_VAULT_CACHE_SIZE, DatabaseManager, type DatabaseManagerLogger, type DatabaseManagerOptions, type DecayCandidateRaw, type DecayClassifier, type DecayInput, type DecayPolicy, type DecaySweepResult, type DecaySweeper, type DecayVerdict, type DurableAutoExtractorOptions, ENTITY_FANOUT, type EmbeddedWalletSignerFn, EncryptionKeyMissingError, type ExtractFactsOptions, type ExtractOutcome, type ExtractedCandidate, type ExtractedEntity, type ExtractionCursorStore, type ExtractionFunnel, type ExtractionTimings, type FactType, type FileMetadata, type FileProcessingReason, type FileProcessingStatus, type FlushResult, type GetConversationsPageOptions, type GetMessagesPageOptions, type GraphTraversalOptions, HARD_DELETE_WINDOW_MS, INFERENCE_ID_HEADER, type InjectionClassifierOptions, type InjectionReason, type LazyStoredConversation, type LlmDecayClassifierOptions, type LlmNeighborRefinerOptions, type LocalMemoryStoreOptions, type Logger, LoggerProvider, type LoggerProviderProps, MAX_HOPS, MEDIUM_TTL_MS, MEMORY_KINDS, MEMORY_LEVELS, type ManualFactType, type MemoriesNeedingTopicExtraction, type MemoryContextItem, type MemoryContextLane, type MemoryContextOptions, type MemoryContextResult, type MemoryCreate, type EmbeddingOptions as MemoryEngineEmbeddingOptions, type MemoryEngineResult, type MemoryEngineSearchOptions, type MemoryExtractedEvent, MemoryKeyError, type MemoryKeyRing, type MemoryKind, type MemoryLevel, MemoryLevelError, type MemoryListOptions, type MemoryMaintenance, type MemoryQuarantinedEvent, type MemoryRecallOptions, type MemoryRetainOptions, type MemoryStore, type MemorySubscribeOptions, type MemoryToVerify, type MemoryTopic, type MemoryUpdate, type MemoryVaultSearchOptions, type MemoryVaultToolOptions, type MemoryVerification, type MessageRedactionResult, type MessageSkeleton, NEVER_TTL_MS, NODE_BUDGET, NOTION_OAUTH_CONFIG, type NeighborRefiner, type NotionAuthUrlParams, type NotionClientRegistration, type NotionExchangeCodeParams, type NotionMcpCaller, type NotionOAuthEndpoints, type NotionPKCEChallenge, type NotionRefreshTokenParams, type NotionTokenResponse, type NowSource, type ObservationTrend, type ObservationTrendInput, PAST_EVENT_GRACE_MS, PII_PATTERNS, PROFILE_DOC_VERSION, type PhotoIngestResult, type PhotoMediaRef, type PiiCategory, type PiiMatch, type PiiPattern, PiiRedactor, type PiiRedactorOptions, type PlatformStorage, type PortalLlmAuth, type PortalLlmFailure, type PortalLlmFailureReason, type ProfileConfigFingerprint, type ProfileDoc, type ProfileFacet, type ProfileFacetKey, type ProfileSalienceInput, type ProfileSection, type PromptPreProcessor, type PromptPreProcessorContext, type PublishedPhotoMemory, type QuantizedEmbedding, type QuarantinedMemoryInfo, QueueManager, type QueueStatus, RECALL_MAX_LIMIT, RECALL_TOOL_NAME, type RankableVaultMemory, type RankedMemory, type RankedProfileCandidate, type RecallContext, type RecallDegradation, type RecallDiagnostics, type RecallEmptyReason, type RecallOptions, type RecallResult, type RecallToolCallbacks, type RecallToolOptions, type RecencyOptions, type RedactionResult, type ReflectOptions, type ReflectResult, type RemoteMemoryCandidateOptions, type RemoteMemoryDecodeFailure, RemoteMemoryError, type RemoteMemoryListOptions, type RemoteMemoryMetadataPage, type RemoteMemoryMetadataRecord, type RemoteMemoryPage, RemoteMemoryPartialCreateError, type RemoteMemoryPersistence, type RemoteMemoryPersistenceOptions, type RemoteMemoryPipeline, type RemoteMemoryPipelineOptions, type RemoteMemoryQueryOptions, type RemoteMemoryQueryPage, type RemoteMemoryReadFilters, type RemoteMemoryRecord, type RemoteMemoryRow, type RemoteMemoryStoreOptions, type RequestEncryptionKeyOptions, RerankerUnavailableError, type ReservationReceipt, type ResumeStreamOptions, type ResumeStreamResult, type ResumeStreamWithStorageResult, type RetainAction, type RetainContext, type RetainOptions, type RetainResult, type RetainSource, SDK_SCHEMA_VERSION, SHORT_TTL_MS, STREAM_RESUMABLE_HEADER, type ScoreBreakdown, type ScoreProfileSalienceOptions, type ScreenResult, type ScreenedCandidate, type SearchSource, type SendMessageWithStorageArgs, type SendMessageWithStorageDetachedResult, type SendMessageWithStorageResult, type ServerToolsOptions, type ServerToolsResponse, type SignMessageFn, type StockPriceClassification, type StockPricePreProcessorOptions, type ChatCompletionUsage as StoredChatCompletionUsage, type StoredConversation, type StoredConversationMemory, type StoredMessage, type StoredMessageWithSimilarity, type StoredVaultFolder, VaultFolder as StoredVaultFolderModel, type StoredVaultMemory, VaultMemory as StoredVaultMemoryModel, StreamExpiredError, type StreamMetaEvent, type StreamResumeHandle, type StreamSmoothingConfig, type SynthesizeProfileOptions, TOOL_RESULT_ORIGIN, TOPICS_EXTRACTION_VERSION, TOPIC_EXTRACTION_BATCH_SIZE, TREND_RECENT_WINDOW_DAYS, TREND_STALE_WINDOW_DAYS, type TextChunk, type TopicExtractOptions, type TopicExtractionInput, type TopicExtractionRunResult, type TopicSkipReason, type TurnCompleteEvent, type TurnSkippedEvent, type UncheckedReason, type UnverifiableReason, type UpdateVaultFolderOptions, type UpdateVaultMemoryOptions, type UseChatStorageOptions, type UseChatStorageResult, type UseCreditsOptions, type UseCreditsResult, type UseModelsOptions, type UseModelsResult, VAULT_SIZE_HOP_CAP, type VaultEmbeddingCache, type VaultEmbeddingExpectation, type VaultFolderOperationsContext, type VaultMemoryKind, type VaultMemoryOperationsContext, type VaultMemoryProjection, type VaultMemoryVisibility, type VaultMemoryWriter, type VaultSaveOperation, type VaultSearchResult, type VaultWriteAction, type VaultWriteInput, type VaultWriteOutcome, type VerificationSources, type VerifyMemoriesForPublishOptions, WalletPoller, type WeatherClassification, type WeatherPreProcessorOptions, type WebSearchClassification, type WebSearchPreProcessorOptions, addConversationMemoriesOp, archiveVaultMemoryOp, assembleMemoryContext, attachFileContextToLastUserMessage, backfillMemoryTopicsOp, buildAttachedFilesText, buildNotionAuthUrl, capHopsForDensity, chatStorageMigrations, chatStorageSchema, chunkAndEmbedAllMessages, chunkAndEmbedMessage, chunkText, classifyCryptoPrice, classifyCryptoPriceBatch, classifyDecay, classifyInjectionCandidates, classifyObservationTrend, classifyStockPrice, classifyStockPriceBatch, classifyWeather, classifyWeatherBatch, classifyWebSearch, classifyWebSearchBatch, clearAllEncryptionKeys, clearAllEncryptionState, clearConversationMemoriesOp, clearEncryptionKey, clearLazyTitleCache, clearServerToolsCache, consoleLogger, conversationMemoryToStored, cosineInt8, createAutoExtractor, createChunkVectorCache, createCryptoPricePreProcessor, createDecaySweeper, createDurableAutoExtractor, createLlmDecayClassifier, createLlmNeighborRefiner, createLocalMemoryStore, createMemoryEngineTool, createMemoryVaultSearchTool, createMemoryVaultTool, createMessageSourceResolver, createNotionProxyTools, createNotionTools, createPlatformCursorStore, createRecallTool, createRemoteMemoryPersistence, createRemoteMemoryPipeline, createRemoteMemoryStore, createStockPricePreProcessor, createStreamingDeAnonymizer, createVaultEmbeddingCache, createVaultFolderOp, createVaultMemoriesBatchOp, createVaultMemoryOp, createWeatherPreProcessor, createWebSearchPreProcessor, decodeChunkVector, decryptConversationTitle, deleteAllVaultMemoriesForUserOp, deleteMessageOp, deleteVaultFolderOp, deleteVaultMemoryOp, dequantizeEmbedding, deriveKeyFromSignatureBytes, deriveMemoryKeyRing, discoverNotionOAuthEndpoints, eagerEmbedContent, embedAllMessages, embedMessage, encodeChunkVector, ensureDefaultFoldersOp, exchangeNotionCode, extractAndLinkEntitiesForMemoriesOp, extractAndRetain, extractEntitiesForMemories, extractFacts, extractReservationReceipts, formatFileProcessingNotes, generateConversationId, generateEmbedding, generateEmbeddings, generateNotionPKCE, getAllVaultFoldersOp, getAllVaultMemoriesOp, getAllVaultMemoryContentsOp, getCachedServerTools, getConversationMemoriesOp, getConversationsByProjectLazyOp, getConversationsLazyOp, getConversationsPageOp, getDecayCandidatesRawOp, getLogger, getMemoriesNeedingTopicExtractionOp, getServerTools, getUnfiledVaultMemoriesOp, getVaultFolderMemoryCountOp, getVaultMemoriesByIdsOp, getVaultMemoryOp, getVaultRankingProjectionsOp, hardDeleteDecayedOp, hasEncryptionKey, ingestPublishedPhotoMemoriesOp, injectionSignatureCatalog, isAttachedFilesText, isDegradedTopicSkip, isEncrypted, isPiiRedactor, isRerankerAvailable, maskScopedEmbeddingCache, memoryCipher, moveMemoriesToFolderOp, noopLogger, onKeyAvailable, parseMedia, preEmbedVaultMemories, quantizeEmbedding, queueManager, rankProfileCandidates, recall, reencryptMemoryField, reflect, refreshEncryptionKeyIfMatches, refreshNotionAccessToken, registerNotionClient, relinkMemoryTopicsOp, requestEncryptionKey, resolvePiiRedactor, restoreVaultMemoryOp, resumeStream, retain, scoreProfileSalience, screenCandidatesForInjection, sdkMigrations, sdkModelClasses, sdkSchema, searchVaultMemories, seedEncryptionKeys, setLogger, setMemoryVisibilityOp, shouldChunkMessage, shouldRecallMemory, stampTopicsExtractedAtOp, streamCancelPath, streamReplayPath, stripLegacyChunkTextOp, summarizeObservationTrends, supersedeVaultMemoryOp, synthesizeProfile, traverseGraphLane, ttlForType, updateVaultFolderContextOp, updateVaultFolderOp, updateVaultMemoryEmbeddingOp, updateVaultMemoryOp, upsertMessageOp, useChat, useChatStorage, useCredits, useEncryption, useModels, verifyMemoriesForPublish, xhrTransport };
+export { type AutoExtractMessage, type AutoExtractor, type Budget, CHUNKS_DISCARDED_ORIGIN, type CachedChunkVectors, type CachedServerTools, Conversation as ChatConversation, Message as ChatMessage, type ChatRole, type ChunkVectorCache, type ChunkingOptions, type ConsolidationAction, type ConsolidationFallbackReason, type ConversationMemoryInput, ConversationMemory as ConversationMemoryModel, type ConversationMemoryOperationsContext, type CreateAutoExtractorOptions, type CreateConversationOptions, type CreateDecaySweeperOptions, type CreateMessageOptions, type CreateVaultFolderOptions, type CreateVaultMemoryOptions, type CryptoPriceClassification, type CryptoPricePreProcessorOptions, DEFAULT_CACHE_EXPIRATION_MS, DEFAULT_CHUNK_CACHE_SIZE, DEFAULT_CHUNK_OVERLAP, DEFAULT_CHUNK_SIZE, DEFAULT_DECAY_POLICY, DEFAULT_MAX_CLASSIFIER_CALLS_PER_SWEEP, DEFAULT_MIN_CHUNK_SIZE, DEFAULT_PROFILE_FACETS, DEFAULT_PROFILE_FACT_TYPE_WEIGHTS, DEFAULT_PROFILE_PROOF_ALPHA, DEFAULT_PROFILE_TREND_MULTIPLIERS, DEFAULT_VAULT_CACHE_SIZE, DatabaseManager, type DatabaseManagerLogger, type DatabaseManagerOptions, type DecayCandidateRaw, type DecayClassifier, type DecayInput, type DecayPolicy, type DecaySweepResult, type DecaySweeper, type DecayVerdict, type DurableAutoExtractorOptions, ENTITY_FANOUT, type EmbeddedWalletSignerFn, EncryptionKeyMissingError, type ExtractFactsOptions, type ExtractOutcome, type ExtractedCandidate, type ExtractedEntity, type ExtractionCursorStore, type ExtractionFunnel, type ExtractionTimings, type FactType, type FileMetadata, type FileProcessingReason, type FileProcessingStatus, type FlushResult, type GetConversationsPageOptions, type GetMessagesPageOptions, type GraphTraversalOptions, HARD_DELETE_WINDOW_MS, INFERENCE_ID_HEADER, type InjectionClassifierOptions, type InjectionReason, type LazyStoredConversation, type LlmDecayClassifierOptions, type LlmNeighborRefinerOptions, type LocalMemoryStoreOptions, type Logger, LoggerProvider, type LoggerProviderProps, MAX_HOPS, MEDIUM_TTL_MS, MEMORY_KINDS, MEMORY_LEVELS, type ManualFactType, type MemoriesNeedingTopicExtraction, type MemoryContextItem, type MemoryContextLane, type MemoryContextOptions, type MemoryContextResult, type MemoryCreate, type EmbeddingOptions as MemoryEngineEmbeddingOptions, type MemoryEngineResult, type MemoryEngineSearchOptions, type MemoryExtractedEvent, MemoryKeyError, type MemoryKeyRing, type MemoryKind, type MemoryLevel, MemoryLevelError, type MemoryListOptions, type MemoryMaintenance, type MemoryQuarantinedEvent, type MemoryRecallOptions, type MemoryRetainOptions, type MemoryStore, type MemorySubscribeOptions, type MemoryToVerify, type MemoryTopic, type MemoryUpdate, type MemoryVaultSearchOptions, type MemoryVaultToolOptions, type MemoryVerification, type MessageRedactionResult, type MessageSkeleton, NEVER_TTL_MS, NODE_BUDGET, NOTION_OAUTH_CONFIG, type NeighborRefiner, type NotionAuthUrlParams, type NotionClientRegistration, type NotionExchangeCodeParams, type NotionMcpCaller, type NotionOAuthEndpoints, type NotionPKCEChallenge, type NotionRefreshTokenParams, type NotionTokenResponse, type NowSource, type ObservationTrend, type ObservationTrendInput, PAST_EVENT_GRACE_MS, PII_PATTERNS, PROFILE_DOC_VERSION, type PhotoIngestResult, type PhotoMediaRef, type PiiCategory, type PiiMatch, type PiiPattern, PiiRedactor, type PiiRedactorOptions, type PlatformStorage, type PortalLlmAuth, type PortalLlmFailure, type PortalLlmFailureReason, type ProfileConfigFingerprint, type ProfileDoc, type ProfileFacet, type ProfileFacetKey, type ProfileSalienceInput, type ProfileSection, type PromptPreProcessor, type PromptPreProcessorContext, type PublishedPhotoMemory, type QuantizedEmbedding, type QuarantinedMemoryInfo, QueueManager, type QueueStatus, RECALL_MAX_LIMIT, RECALL_TOOL_NAME, type RankableVaultMemory, type RankedMemory, type RankedProfileCandidate, type RecallContext, type RecallDegradation, type RecallDiagnostics, type RecallEmptyReason, type RecallOptions, type RecallResult, type RecallToolCallbacks, type RecallToolOptions, type RecencyOptions, type RedactionResult, type ReflectOptions, type ReflectResult, type RemoteMemoryCandidateOptions, type RemoteMemoryDecodeFailure, RemoteMemoryError, type RemoteMemoryListOptions, type RemoteMemoryMetadataPage, type RemoteMemoryMetadataRecord, type RemoteMemoryPage, RemoteMemoryPartialCreateError, type RemoteMemoryPersistence, type RemoteMemoryPersistenceOptions, type RemoteMemoryPipeline, type RemoteMemoryPipelineOptions, type RemoteMemoryQueryOptions, type RemoteMemoryQueryPage, type RemoteMemoryReadFilters, type RemoteMemoryRecord, type RemoteMemoryRow, type RemoteMemoryStoreOptions, type RequestEncryptionKeyOptions, RerankerUnavailableError, type ReservationReceipt, type ResumeStreamOptions, type ResumeStreamResult, type ResumeStreamWithStorageResult, type RetainAction, type RetainContext, type RetainOptions, type RetainResult, type RetainSource, SDK_SCHEMA_VERSION, SHORT_TTL_MS, STREAM_RESUMABLE_HEADER, type ScoreBreakdown, type ScoreProfileSalienceOptions, type ScreenResult, type ScreenedCandidate, type SearchSource, type SendMessageWithStorageArgs, type SendMessageWithStorageDetachedResult, type SendMessageWithStorageResult, type ServerToolsOptions, type ServerToolsResponse, type SignMessageFn, type StockPriceClassification, type StockPricePreProcessorOptions, type ChatCompletionUsage as StoredChatCompletionUsage, type StoredConversation, type StoredConversationMemory, type StoredMessage, type StoredMessageWithSimilarity, type StoredVaultFolder, VaultFolder as StoredVaultFolderModel, type StoredVaultMemory, VaultMemory as StoredVaultMemoryModel, StreamExpiredError, type StreamMetaEvent, type StreamResumeHandle, type StreamSmoothingConfig, type SynthesizeProfileOptions, TOOL_RESULT_ORIGIN, TOPICS_EXTRACTION_VERSION, TOPIC_EXTRACTION_BATCH_SIZE, TREND_RECENT_WINDOW_DAYS, TREND_STALE_WINDOW_DAYS, type TextChunk, type TopicExtractOptions, type TopicExtractionInput, type TopicExtractionRunResult, type TopicSkipReason, type TurnCompleteEvent, type TurnSkippedEvent, type UncheckedReason, type UnverifiableReason, type UpdateVaultFolderOptions, type UpdateVaultMemoryOptions, type UseChatStorageOptions, type UseChatStorageResult, type UseCreditsOptions, type UseCreditsResult, type UseModelsOptions, type UseModelsResult, VAULT_SIZE_HOP_CAP, type VaultEmbeddingCache, type VaultEmbeddingExpectation, type VaultFolderOperationsContext, type VaultMemoryKind, type VaultMemoryOperationsContext, type VaultMemoryProjection, type VaultMemoryVisibility, type VaultMemoryWriter, type VaultSaveOperation, type VaultSearchResult, type VaultWriteAction, type VaultWriteInput, type VaultWriteOutcome, type VerificationSources, type VerifyMemoriesForPublishOptions, WalletPoller, type WeatherClassification, type WeatherPreProcessorOptions, type WebSearchClassification, type WebSearchPreProcessorOptions, addConversationMemoriesOp, archiveVaultMemoryOp, assembleMemoryContext, attachFileContextToLastUserMessage, backfillMemoryTopicsOp, buildAttachedFilesText, buildNotionAuthUrl, capHopsForDensity, chatStorageMigrations, chatStorageSchema, chunkAndEmbedAllMessages, chunkAndEmbedMessage, chunkText, classifyCryptoPrice, classifyCryptoPriceBatch, classifyDecay, classifyInjectionCandidates, classifyObservationTrend, classifyStockPrice, classifyStockPriceBatch, classifyWeather, classifyWeatherBatch, classifyWebSearch, classifyWebSearchBatch, clearAllEncryptionKeys, clearAllEncryptionState, clearConversationMemoriesOp, clearEncryptionKey, clearLazyTitleCache, clearModelsCache, clearServerToolsCache, consoleLogger, conversationMemoryToStored, cosineInt8, createAutoExtractor, createChunkVectorCache, createCryptoPricePreProcessor, createDecaySweeper, createDurableAutoExtractor, createLlmDecayClassifier, createLlmNeighborRefiner, createLocalMemoryStore, createMemoryEngineTool, createMemoryVaultSearchTool, createMemoryVaultTool, createMessageSourceResolver, createNotionProxyTools, createNotionTools, createPlatformCursorStore, createRecallTool, createRemoteMemoryPersistence, createRemoteMemoryPipeline, createRemoteMemoryStore, createStockPricePreProcessor, createStreamingDeAnonymizer, createVaultEmbeddingCache, createVaultFolderOp, createVaultMemoriesBatchOp, createVaultMemoryOp, createWeatherPreProcessor, createWebSearchPreProcessor, decodeChunkVector, decryptConversationTitle, deleteAllVaultMemoriesForUserOp, deleteMessageOp, deleteVaultFolderOp, deleteVaultMemoryOp, dequantizeEmbedding, deriveKeyFromSignatureBytes, deriveMemoryKeyRing, discoverNotionOAuthEndpoints, eagerEmbedContent, embedAllMessages, embedMessage, encodeChunkVector, ensureDefaultFoldersOp, exchangeNotionCode, extractAndLinkEntitiesForMemoriesOp, extractAndRetain, extractEntitiesForMemories, extractFacts, extractReservationReceipts, formatFileProcessingNotes, generateConversationId, generateEmbedding, generateEmbeddings, generateNotionPKCE, getAllVaultFoldersOp, getAllVaultMemoriesOp, getAllVaultMemoryContentsOp, getCachedServerTools, getConversationMemoriesOp, getConversationsByProjectLazyOp, getConversationsLazyOp, getConversationsPageOp, getDecayCandidatesRawOp, getLogger, getMemoriesNeedingTopicExtractionOp, getServerTools, getUnfiledVaultMemoriesOp, getVaultFolderMemoryCountOp, getVaultMemoriesByIdsOp, getVaultMemoryOp, getVaultRankingProjectionsOp, hardDeleteDecayedOp, hasEncryptionKey, ingestPublishedPhotoMemoriesOp, injectionSignatureCatalog, isAttachedFilesText, isDegradedTopicSkip, isEncrypted, isPiiRedactor, isRerankerAvailable, maskScopedEmbeddingCache, memoryCipher, moveMemoriesToFolderOp, noopLogger, onKeyAvailable, parseMedia, preEmbedVaultMemories, quantizeEmbedding, queueManager, rankProfileCandidates, recall, reencryptMemoryField, reflect, refreshEncryptionKeyIfMatches, refreshNotionAccessToken, registerNotionClient, relinkMemoryTopicsOp, requestEncryptionKey, resolvePiiRedactor, restoreVaultMemoryOp, resumeStream, retain, scoreProfileSalience, screenCandidatesForInjection, sdkMigrations, sdkModelClasses, sdkSchema, searchVaultMemories, seedEncryptionKeys, setLogger, setMemoryVisibilityOp, shouldChunkMessage, shouldRecallMemory, stampTopicsExtractedAtOp, streamCancelPath, streamReplayPath, stripLegacyChunkTextOp, summarizeObservationTrends, supersedeVaultMemoryOp, synthesizeProfile, traverseGraphLane, ttlForType, updateVaultFolderContextOp, updateVaultFolderOp, updateVaultMemoryEmbeddingOp, updateVaultMemoryOp, upsertMessageOp, useChat, useChatStorage, useCredits, useEncryption, useModels, verifyMemoriesForPublish, xhrTransport };

package/dist/expo/index.mjs

@@ -4296,15 +4296,17 @@ Build ONLY when the user gives an explicit instruction to build, create, generat
 
 Once the user HAS given an explicit build instruction, you are in App Builder mode: you produce polished, production-quality React apps that feel designed \u2014 not generic.
 
-DESIGN: on the initial build, before writing code, briefly state (2-3 sentences) what you want this app to feel like \u2014 name the specific fonts, the palette in hex, and one signature detail. Be concrete: "Fraunces display + Inter body + JetBrains Mono metadata, terracotta #b75432 accent on bone #f3efe6, paper-grain texture via inline SVG noise" is the kind of specificity that anchors the build. Default to nothing: don't pick "safe modern productivity" unless the brief literally asks for utilitarian. On later turns, state a design brief only when the change is visual (a restyle, new theme, or new UI); for a logic, data, or behavior-only change, skip the brief and just make the change.
+DESIGN: on the initial build, before writing code, briefly state (2-3 sentences) what you want this app to feel like \u2014 name the specific fonts, the palette in hex, and one signature detail. Be concrete: "Fraunces display + Inter body + JetBrains Mono metadata, terracotta #b75432 accent on bone #f3efe6, paper-grain texture via inline SVG noise" is the kind of specificity that anchors the build. Default to nothing: don't pick "safe modern productivity" unless the brief literally asks for utilitarian. On later turns, state a design brief only when the change is visual (a restyle, new theme, or new UI); for a text, logic, data, or behavior-only change, skip the brief and just make the change.
+
+EDIT SCOPE: Change only the files and lines required by the current user request. A title rename is a text edit, not a design change. Preserve styles, dependencies, and unrelated logic on text edits. Design feedback does not authorize unrelated edits.
 
 WORKFLOW:
 1. NEVER output code as text or markdown. ALWAYS use tools.
 2. Initial build: state your design brief, then create_file with a "files" array writing ALL files in one call. The preview renders automatically.
 3. patch_file requires read_file for that path earlier in the conversation (or create_file in the same conversation, which counts). read_file output is numbered "42: <text>" \u2014 the prefix is display-only; never copy it into "find". Each "find" must match exactly one location \u2014 include 2-3 lines of surrounding context. Pass multiple patches per call to change distinct locations.
 4. Prefer patch_file for changes to existing files \u2014 surgical, reviewable, preserves surrounding code. Reserve create_file overwrites for substantial restructuring. Inserts: find the line before, replace with that line + the new line. Deletes: empty replace.
-5. After substantial changes, call critique_design first (taste, hierarchy, distinctiveness, weakest decision) and then audit_design (mechanical: tokens, focus states, aria-labels, alt, heading order). Patch what each surfaces. critique catches taste issues no checklist would; audit catches mechanics you'd miss. Run both; act on both.
-6. Before declaring done, call verify_app. The host's preview runtime reports whether your app actually mounted and surfaces any errors it captured \u2014 hallucinated imports, syntax the bundler rejected, undefined components, runtime crashes. If \`rendered\` is false or \`errors\` is non-empty, fix every error before the next message. If the response includes a \`note\` that verify_app isn't wired in this host, skip \u2014 audit + critique are your feedback in that case.
+5. On the initial build, when adding new UI, or after user-requested design changes, call critique_design first and then audit_design. Critique checks taste, hierarchy, and distinctiveness. Audit checks tokens, focus states, aria-labels, alt text, and heading order. Patch only issues within the current user request. For text, logic, or data edits, skip both design tools.
+6. Before declaring done, call verify_app. The host's preview runtime reports whether your app actually mounted and surfaces any errors it captured \u2014 hallucinated imports, syntax the bundler rejected, undefined components, runtime crashes. If \`rendered\` is false or \`errors\` is non-empty, fix every error before the next message. If the response includes a \`note\` that verify_app isn't wired in this host, skip runtime verification. An unavailable runtime verifier does not authorize design changes. Use audit_design and critique_design only for the initial build, adding new UI, or user-requested design changes.
 7. Keep responses to a sentence or two. Exception: when responding to critique_design, write a substantive paragraph per rubric question.
 
 STRUCTURE:
@@ -4470,12 +4472,13 @@ function isCacheExpired(cache, expirationMs = DEFAULT_CACHE_EXPIRATION_MS) {
   if (!cache) return true;
   return Date.now() - cache.timestamp > expirationMs;
 }
-function buildCacheEntry(tools, checksum) {
+function buildCacheEntry(tools, checksum, etag) {
   return {
     tools,
     timestamp: Date.now(),
     version: CACHE_VERSION,
-    ...checksum && { checksum }
+    ...checksum && { checksum },
+    ...etag && { etag }
   };
 }
 function writeLocalStorageCache(entry) {
@@ -4510,19 +4513,36 @@ function shouldRefreshTools(responseChecksum, cache = localStorageToolsCache) {
   const cachedChecksum = getToolsChecksum2(cache);
   return cachedChecksum instanceof Promise ? cachedChecksum.then(decide) : decide(cachedChecksum);
 }
-async function fetchServerToolsFromApi(baseUrl, token) {
-  const response = await fetch(`${baseUrl}/api/v1/tools`, {
+async function fetchServerToolsFromApi(baseUrl, authHeaders, etag) {
+  const request = (validator) => fetch(`${baseUrl}/api/v1/tools`, {
     method: "GET",
     headers: {
-      Authorization: `Bearer ${token}`,
-      "Content-Type": "application/json"
+      ...authHeaders,
+      "Content-Type": "application/json",
+      ...validator && { "If-None-Match": validator }
     }
   });
+  let response;
+  try {
+    response = await request(etag);
+  } catch (error) {
+    if (!etag || !(error instanceof TypeError)) throw error;
+    etag = void 0;
+    response = await request();
+  }
+  if (etag && response.status === 304) {
+    return { notModified: true };
+  }
   if (!response.ok) {
     throw new Error(`Failed to fetch server tools: ${response.status}`);
   }
   const data = await response.json();
-  return convertServerToolsResponse(data);
+  const responseEtag = response.headers?.get("ETag") ?? void 0;
+  return {
+    notModified: false,
+    ...convertServerToolsResponse(data),
+    ...responseEtag && { etag: responseEtag }
+  };
 }
 async function getServerTools(options) {
   const {
@@ -4554,34 +4574,30 @@ async function getServerTools(options) {
   try {
     const { BASE_URL: BASE_URL2 } = await import("./clientConfig-VPOGQRNP.mjs");
     const effectiveBaseUrl = baseUrl ?? BASE_URL2;
+    const validator = cached && !forceRefresh ? cached.etag : void 0;
+    let authHeaders;
     if (apiKey) {
-      const response = await fetch(`${effectiveBaseUrl}/api/v1/tools`, {
-        method: "GET",
-        headers: {
-          "X-API-Key": apiKey,
-          "Content-Type": "application/json"
-        }
-      });
-      if (!response.ok) {
-        throw new Error(`Failed to fetch server tools: ${response.status}`);
+      authHeaders = { "X-API-Key": apiKey };
+    } else {
+      if (!getToken) {
+        getLogger().warn("[serverTools] No auth method available for fetching tools");
+        return cached?.tools ?? [];
       }
-      const data = await response.json();
-      const { tools: tools2, checksum: checksum2 } = convertServerToolsResponse(data);
-      await persistCache(buildCacheEntry(tools2, checksum2));
-      return tools2;
-    }
-    if (!getToken) {
-      getLogger().warn("[serverTools] No auth method available for fetching tools");
-      return cached?.tools ?? [];
+      const token = await getToken();
+      if (!token) {
+        getLogger().warn("[serverTools] No auth token available for fetching tools");
+        return cached?.tools ?? [];
+      }
+      authHeaders = { Authorization: `Bearer ${token}` };
     }
-    const token = await getToken();
-    if (!token) {
-      getLogger().warn("[serverTools] No auth token available for fetching tools");
-      return cached?.tools ?? [];
+    const result = await fetchServerToolsFromApi(effectiveBaseUrl, authHeaders, validator);
+    if (result.notModified) {
+      const confirmed = cached;
+      await persistCache({ ...confirmed, timestamp: Date.now() });
+      return confirmed.tools;
     }
-    const { tools, checksum } = await fetchServerToolsFromApi(effectiveBaseUrl, token);
-    await persistCache(buildCacheEntry(tools, checksum));
-    return tools;
+    await persistCache(buildCacheEntry(result.tools, result.checksum, result.etag));
+    return result.tools;
   } catch (error) {
     getLogger().error("[serverTools] Failed to fetch server tools:", error);
     if (cached?.tools) {
@@ -22032,6 +22048,128 @@ function useCredits(options = {}) {
 
 // src/react/useModels.ts
 import { useCallback as useCallback2, useEffect as useEffect2, useRef as useRef2, useState as useState2 } from "react";
+var MODELS_CACHE_TTL_MS = 5 * 60 * 1e3;
+var modelsCache = /* @__PURE__ */ new Map();
+var modelsInFlight = /* @__PURE__ */ new Map();
+var cacheGeneration = 0;
+function clearModelsCache() {
+  cacheGeneration++;
+  modelsCache.clear();
+  modelsInFlight.clear();
+}
+function pruneStaleTokenEntries(newKey) {
+  const [newBaseUrl, newProvider] = JSON.parse(newKey);
+  for (const existingKey of modelsCache.keys()) {
+    if (existingKey === newKey) continue;
+    const [existingBaseUrl, existingProvider] = JSON.parse(existingKey);
+    if (existingBaseUrl === newBaseUrl && existingProvider === newProvider) {
+      modelsCache.delete(existingKey);
+    }
+  }
+}
+function fingerprintToken(token) {
+  if (!token) return "anon";
+  let h1 = 3735928559;
+  let h2 = 1103547991;
+  for (let i = 0; i < token.length; i++) {
+    const ch = token.charCodeAt(i);
+    h1 = Math.imul(h1 ^ ch, 2654435761);
+    h2 = Math.imul(h2 ^ ch, 1597334677);
+  }
+  h1 = Math.imul(h1 ^ h1 >>> 16, 2246822507) ^ Math.imul(h2 ^ h2 >>> 13, 3266489909);
+  h2 = Math.imul(h2 ^ h2 >>> 16, 2246822507) ^ Math.imul(h1 ^ h1 >>> 13, 3266489909);
+  return (4294967296 * (2097151 & h2) + (h1 >>> 0)).toString(36);
+}
+function modelsCacheKey(baseUrl, provider, token) {
+  return JSON.stringify([baseUrl, provider ?? null, fingerprintToken(token)]);
+}
+async function fetchAllModelPages(baseUrl, provider, headers, previous, signal) {
+  const pages = /* @__PURE__ */ new Map();
+  const models = [];
+  let pageToken;
+  do {
+    const pageKey = pageToken ?? "";
+    const stored2 = previous?.pages.get(pageKey);
+    const request = (etag) => getApiV1Models({
+      baseUrl,
+      headers: etag ? { ...headers, "If-None-Match": etag } : headers,
+      query: { provider, page_token: pageToken },
+      signal
+    });
+    let response = await request(stored2?.etag);
+    if (stored2?.etag && !response.response && response.error) {
+      response = await request();
+    }
+    let page;
+    if (response.response?.status === 304) {
+      if (!stored2) throw new Error("Failed to fetch models");
+      page = stored2;
+    } else {
+      if (response.error) {
+        const errorMsg = response.error.error ?? "Failed to fetch models";
+        throw new Error(errorMsg);
+      }
+      page = {
+        body: response.data ?? {},
+        etag: response.response?.headers?.get("ETag") ?? void 0
+      };
+    }
+    pages.set(pageKey, page);
+    models.push(...page.body.data || []);
+    pageToken = page.body.next_page_token;
+  } while (pageToken);
+  return { pages, models, fetchedAt: Date.now() };
+}
+function loadModels(key, baseUrl, provider, headers, force, signal) {
+  const cached = modelsCache.get(key);
+  if (cached && !force && Date.now() - cached.fetchedAt < MODELS_CACHE_TTL_MS) {
+    return Promise.resolve([...cached.models]);
+  }
+  const shared = (force ? void 0 : modelsInFlight.get(key)) ?? startSharedRequest(key, baseUrl, provider, headers, cached);
+  return waitForSharedRequest(key, shared, signal).then((models) => [...models]);
+}
+function startSharedRequest(key, baseUrl, provider, headers, previous) {
+  const controller = new AbortController();
+  const generation = cacheGeneration;
+  const shared = {
+    controller,
+    waiters: 0,
+    promise: Promise.resolve([])
+  };
+  shared.promise = fetchAllModelPages(baseUrl, provider, headers, previous, controller.signal).then((entry) => {
+    if (modelsInFlight.get(key) === shared && generation === cacheGeneration) {
+      pruneStaleTokenEntries(key);
+      modelsCache.set(key, entry);
+    }
+    return entry.models;
+  }).finally(() => {
+    if (modelsInFlight.get(key) === shared) modelsInFlight.delete(key);
+  });
+  modelsInFlight.set(key, shared);
+  return shared;
+}
+function waitForSharedRequest(key, shared, signal) {
+  shared.waiters++;
+  let left = false;
+  const leave = () => {
+    if (left) return;
+    left = true;
+    shared.waiters--;
+    if (shared.waiters === 0) {
+      shared.controller.abort();
+      if (modelsInFlight.get(key) === shared) modelsInFlight.delete(key);
+    }
+  };
+  if (signal.aborted) {
+    leave();
+  } else {
+    signal.addEventListener("abort", leave, { once: true });
+  }
+  return shared.promise.finally(() => {
+    left = true;
+    signal.removeEventListener("abort", leave);
+  });
+}
 function useModels(options = {}) {
   const { getToken, baseUrl = BASE_URL, provider, autoFetch = true } = options;
   const [models, setModels] = useState2([]);
@@ -22054,7 +22192,7 @@ function useModels(options = {}) {
       }
     };
   }, []);
-  const fetchModels = useCallback2(async () => {
+  const fetchModels = useCallback2(async (force = false) => {
     if (abortControllerRef.current) {
       abortControllerRef.current.abort();
     }
@@ -22073,35 +22211,17 @@ function useModels(options = {}) {
       if (token) {
         headers["Authorization"] = `Bearer ${token}`;
       }
-      let allModels = [];
-      let nextPageToken;
-      do {
-        if (signal.aborted) return;
-        const response = await getApiV1Models({
-          baseUrl: baseUrlRef.current,
-          headers,
-          query: {
-            provider: providerRef.current,
-            page_token: nextPageToken
-          },
-          signal
-        });
-        if (response.error) {
-          const errorMsg = response.error.error ?? "Failed to fetch models";
-          throw new Error(errorMsg);
-        }
-        if (response.data) {
-          const newModels = response.data.data || [];
-          allModels = [...allModels, ...newModels];
-          nextPageToken = response.data.next_page_token;
-        }
-      } while (nextPageToken);
+      const baseUrl2 = baseUrlRef.current;
+      const provider2 = providerRef.current;
+      const key = modelsCacheKey(baseUrl2, provider2, token);
+      const allModels = await loadModels(key, baseUrl2, provider2, headers, force, signal);
       if (signal.aborted) return;
       setModels(allModels);
     } catch (err) {
       if (err instanceof Error && err.name === "AbortError") {
         return;
       }
+      if (signal.aborted) return;
       setError(err instanceof Error ? err : new Error(String(err)));
     } finally {
       if (!signal.aborted) {
@@ -22114,7 +22234,7 @@ function useModels(options = {}) {
   }, []);
   const refetch = useCallback2(async () => {
     setModels([]);
-    await fetchModels();
+    await fetchModels(true);
   }, [fetchModels]);
   const hasFetchedRef = useRef2(false);
   useEffect2(() => {
@@ -22840,6 +22960,7 @@ function useChat(options) {
     apiType: defaultApiType = "auto",
     smoothing,
     preProcessors,
+    hooks,
     piiRedaction,
     onPiiRedacted,
     resumable = false,
@@ -23093,6 +23214,7 @@ function useChat(options) {
           onToolCallArgumentsDelta,
           onStepFinish,
           preProcessors,
+          hooks,
           piiRedaction: requestPiiRedaction ?? resolvedPiiRedaction,
           onPiiRedacted
         });
@@ -23132,6 +23254,7 @@ function useChat(options) {
       defaultApiType,
       smoothing,
       preProcessors,
+      hooks,
       resolvedPiiRedaction,
       onPiiRedacted,
       resumable,
@@ -24092,6 +24215,7 @@ function useChatStorage(options) {
     embeddingModel = DEFAULT_API_EMBEDDING_MODEL,
     minContentLength = DEFAULT_MIN_CONTENT_LENGTH,
     preProcessors,
+    hooks,
     smoothing,
     resumable = false,
     onCancelResult,
@@ -24551,6 +24675,7 @@ function useChatStorage(options) {
     onError,
     apiType,
     preProcessors,
+    hooks,
     smoothing,
     resumable,
     onCancelResult,
@@ -25899,6 +26024,7 @@ export {
   clearConversationMemoriesOp,
   clearEncryptionKey,
   clearLazyTitleCache,
+  clearModelsCache,
   clearServerToolsCache,
   consoleLogger,
   conversationMemoryToStored,