@mastra/libsql 1.7.0-alpha.0 → 1.7.1-alpha.0

package/dist/docs/references/docs-observability-overview.md

@@ -1,8 +1,8 @@
-# Observability Overview
+# Observability overview
 
 Mastra provides observability features for AI applications. Monitor LLM operations, trace agent decisions, and debug complex workflows with tools that understand AI-specific patterns.
 
-## Key Features
+## Key features
 
 ### Tracing
 
@@ -13,13 +13,13 @@ Specialized tracing for AI operations that captures:
 - **Workflow steps**: Branching logic, parallel execution, and step outputs
 - **Automatic instrumentation**: Tracing with decorators
 
-## Storage Requirements
+## Storage requirements
 
 The `DefaultExporter` persists traces to your configured storage backend. Not all storage providers support observability—for the full list, see [Storage Provider Support](https://mastra.ai/docs/observability/tracing/exporters/default).
 
 For production environments with high traffic, we recommend using **ClickHouse** for the observability domain via [composite storage](https://mastra.ai/reference/storage/composite). See [Production Recommendations](https://mastra.ai/docs/observability/tracing/exporters/default) for details.
 
-## Quick Start
+## Quick start
 
 Configure Observability in your Mastra instance:
 
@@ -63,7 +63,7 @@ With this basic setup, you will see Traces and Logs in both Studio and in Mastra
 
 We also support various external tracing providers like MLflow, Langfuse, Braintrust, and any OpenTelemetry-compatible platform (Datadog, New Relic, SigNoz, etc.). See more about this in the [Tracing](https://mastra.ai/docs/observability/tracing/overview) documentation.
 
-## What's Next?
+## What's next?
 
 - **[Set up Tracing](https://mastra.ai/docs/observability/tracing/overview)**: Configure tracing for your application
 - **[Configure Logging](https://mastra.ai/docs/observability/logging)**: Add structured logging

package/dist/docs/references/docs-observability-tracing-exporters-default.md

@@ -1,4 +1,4 @@
-# Default Exporter
+# Default exporter
 
 The `DefaultExporter` persists traces to your configured storage backend, making them accessible through Studio. It's automatically enabled when using the default observability configuration and requires no external services.
 
@@ -68,7 +68,7 @@ export const mastra = new Mastra({
 })
 ```
 
-## Viewing Traces
+## Viewing traces
 
 ### Studio
 
@@ -79,7 +79,7 @@ Access your traces through Studio:
 3. Filter and search your local traces
 4. Inspect detailed span information
 
-## Tracing Strategies
+## Tracing strategies
 
 DefaultExporter automatically selects the optimal tracing strategy based on your storage provider. You can also override this selection if needed.
 
@@ -106,7 +106,7 @@ new DefaultExporter({
 })
 ```
 
-## Storage Provider Support
+## Storage provider support
 
 Different storage providers support different tracing strategies. Some providers support observability for production workloads, while others are intended primarily for local development.
 
@@ -139,7 +139,7 @@ The following storage providers **don't support** the observability domain. If y
 - **batch-with-updates**: 10-100x throughput improvement, full span lifecycle
 - **insert-only**: Additional 70% reduction in database operations, perfect for analytics
 
-## Production Recommendations
+## Production recommendations
 
 Observability data grows quickly in production environments. A single agent interaction can generate hundreds of spans, and high-traffic applications can produce thousands of traces per day. Most general-purpose databases aren't optimized for this write-heavy, append-only workload.
 
@@ -156,7 +156,7 @@ Observability data grows quickly in production environments. A single agent inte
 
 If you're using a provider without observability support (like Convex or DynamoDB) or want to optimize performance, use [composite storage](https://mastra.ai/reference/storage/composite) to route observability data to ClickHouse while keeping other data in your primary database.
 
-## Batching Behavior
+## Batching behavior
 
 ### Flush Triggers
 
@@ -167,7 +167,7 @@ For both batch strategies (`batch-with-updates` and `insert-only`), traces are f
 3. **Emergency flush**: Buffer approaches `maxBufferSize` limit
 4. **Shutdown**: Force flush all pending events
 
-### Error Handling
+### Error handling
 
 The DefaultExporter includes robust error handling for production use:
 

package/dist/docs/references/docs-rag-retrieval.md

@@ -1,10 +1,10 @@
-# Retrieval in RAG Systems
+# Retrieval in RAG systems
 
 After storing embeddings, you need to retrieve relevant chunks to answer user queries.
 
 Mastra provides flexible retrieval options with support for semantic search, filtering, and re-ranking.
 
-## How Retrieval Works
+## How retrieval works
 
 1. The user's query is converted to an embedding using the same model used for document embeddings
 2. This embedding is compared to stored embeddings using vector similarity
@@ -14,7 +14,7 @@ Mastra provides flexible retrieval options with support for semantic search, fil
 - Re-ranked for better relevance
 - Processed through a knowledge graph
 
-## Basic Retrieval
+## Basic retrieval
 
 The simplest approach is direct semantic search. This method uses vector similarity to find chunks that are semantically similar to the query:
 
@@ -63,7 +63,7 @@ Results include both the text content and a similarity score:
 ]
 ```
 
-## Advanced Retrieval options
+## Advanced retrieval options
 
 ### Metadata Filtering
 
@@ -272,7 +272,7 @@ import { PGVECTOR_PROMPT } from '@mastra/pg'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${PGVECTOR_PROMPT}
@@ -289,7 +289,7 @@ import { PINECONE_PROMPT } from '@mastra/pinecone'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${PINECONE_PROMPT}
@@ -306,7 +306,7 @@ import { QDRANT_PROMPT } from '@mastra/qdrant'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${QDRANT_PROMPT}
@@ -323,7 +323,7 @@ import { CHROMA_PROMPT } from '@mastra/chroma'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${CHROMA_PROMPT}
@@ -340,7 +340,7 @@ import { ASTRA_PROMPT } from '@mastra/astra'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${ASTRA_PROMPT}
@@ -357,7 +357,7 @@ import { LIBSQL_PROMPT } from '@mastra/libsql'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${LIBSQL_PROMPT}
@@ -374,7 +374,7 @@ import { UPSTASH_PROMPT } from '@mastra/upstash'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${UPSTASH_PROMPT}
@@ -391,7 +391,7 @@ import { VECTORIZE_PROMPT } from '@mastra/vectorize'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${VECTORIZE_PROMPT}
@@ -408,7 +408,7 @@ import { MONGODB_PROMPT } from '@mastra/mongodb'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${MONGODB_PROMPT}
@@ -425,7 +425,7 @@ import { OPENSEARCH_PROMPT } from '@mastra/opensearch'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${OPENSEARCH_PROMPT}
@@ -442,7 +442,7 @@ import { S3VECTORS_PROMPT } from '@mastra/s3vectors'
 export const ragAgent = new Agent({
   id: 'rag-agent',
   name: 'RAG Agent',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   instructions: `
   Process queries using the provided context. Structure responses to be concise and relevant.
   ${S3VECTORS_PROMPT}
@@ -472,7 +472,7 @@ const initialResults = await pgVector.query({
 })
 
 // Create a relevance scorer
-const relevanceProvider = new MastraAgentRelevanceScorer('relevance-scorer', 'openai/gpt-5.1')
+const relevanceProvider = new MastraAgentRelevanceScorer('relevance-scorer', 'openai/gpt-5.4')
 
 // Re-rank the results
 const rerankedResults = await rerank({

package/dist/docs/references/reference-core-getMemory.md

@@ -22,7 +22,7 @@ const thread = await memory.createThread({
 
 **memory** (`TMemory[TMemoryKey]`): The memory instance with the specified key. Throws an error if the memory is not found.
 
-## Example: Registering and Retrieving Memory
+## Example: Registering and retrieving memory
 
 ```typescript
 import { Mastra } from '@mastra/core'

package/dist/docs/references/reference-core-listMemory.md

@@ -20,7 +20,7 @@ This method takes no parameters.
 
 **memory** (`Record<string, MastraMemory>`): An object containing all registered memory instances, keyed by their registry keys.
 
-## Example: Checking Registered Memory
+## Example: Checking registered memory
 
 ```typescript
 import { Mastra } from '@mastra/core'

package/dist/docs/references/reference-core-mastra-class.md

@@ -1,4 +1,4 @@
-# Mastra Class
+# Mastra class
 
 The `Mastra` class is the central orchestrator in any Mastra application, managing agents, workflows, storage, logging, observability, and more. Typically, you create a single instance of `Mastra` to coordinate your application.
 
@@ -55,7 +55,7 @@ Visit the [Configuration reference](https://mastra.ai/reference/configuration) f
 
 **mcpServers** (`Record<string, MCPServerBase>`): An object where keys are registry keys (used for getMCPServer()) and values are instances of MCPServer or classes extending MCPServerBase. Each MCPServer must have an id property. Servers can be retrieved by registry key using getMCPServer() or by their intrinsic id using getMCPServerById().
 
-**bundler** (`BundlerConfig`): Configuration for the asset bundler with options for externals, sourcemap, and transpilePackages.
+**bundler** (`BundlerConfig`): Configuration for the asset bundler with options for externals, sourcemap, transpilePackages, and dynamicPackages. (Default: `{ externals: [], sourcemap: false, transpilePackages: [], dynamicPackages: [] }`)
 
 **scorers** (`Record<string, Scorer>`): Scorers for evaluating agent responses and workflow outputs (Default: `{}`)
 

package/dist/docs/references/reference-memory-memory-class.md

@@ -1,4 +1,4 @@
-# Memory Class
+# Memory class
 
 The `Memory` class provides a robust system for managing conversation history and thread-based message storage in Mastra. It enables persistent storage of conversations, semantic search capabilities, and efficient message retrieval. You must configure a storage provider for conversation history, and if you enable semantic recall you will also need to provide a vector store and embedder.
 
@@ -11,7 +11,7 @@ import { Agent } from '@mastra/core/agent'
 export const agent = new Agent({
   name: 'test-agent',
   instructions: 'You are an agent with memory.',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   memory: new Memory({
     options: {
       workingMemory: {
@@ -60,7 +60,7 @@ import { LibSQLStore, LibSQLVector } from '@mastra/libsql'
 export const agent = new Agent({
   name: 'test-agent',
   instructions: 'You are an agent with memory.',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   memory: new Memory({
     storage: new LibSQLStore({
       id: 'test-agent-storage',
@@ -97,7 +97,7 @@ import { PgStore, PgVector } from '@mastra/pg'
 export const agent = new Agent({
   name: 'pg-agent',
   instructions: 'You are an agent with optimized PostgreSQL memory.',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   memory: new Memory({
     storage: new PgStore({
       id: 'pg-agent-storage',

package/dist/docs/references/reference-storage-composite.md

@@ -1,4 +1,4 @@
-# Composite Storage
+# Composite storage
 
 `MastraCompositeStore` can compose storage domains from different providers. Use it when you need different databases for different purposes. For example, use LibSQL for memory and PostgreSQL for workflows.
 
@@ -58,7 +58,7 @@ bun add @mastra/pg@latest @mastra/libsql@latest
 
 ## Storage domains
 
-Mastra organizes storage into five specialized domains, each handling a specific type of data. Each domain can be backed by a different storage adapter, and domain classes are exported from each storage package.
+Mastra organizes storage into domains, each handling a specific type of data. Each domain can be backed by a different storage adapter, and domain classes are exported from each storage package.
 
 | Domain          | Description                                                                                                                                                                           |
 | --------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
@@ -67,6 +67,10 @@ Mastra organizes storage into five specialized domains, each handling a specific
 | `scores`        | Evaluation results from Mastra's evals system. Scores and metrics are persisted here for analysis and comparison over time.                                                           |
 | `observability` | Telemetry data including traces and spans. Agent interactions, tool calls, and LLM requests generate spans collected into traces for debugging and performance analysis.              |
 | `agents`        | Agent configurations for stored agents. Enables agents to be defined and updated at runtime without code deployments.                                                                 |
+| `datasets`      | Evaluation datasets used for experiment runs. Stores dataset definitions, schemas, and versioned items.                                                                               |
+| `experiments`   | Experiment runs and per-item experiment results linked to datasets and targets.                                                                                                       |
+
+> **Note:** `MastraCompositeStore` accepts all of the domain keys above, but storage adapter support varies by package. You can mix adapters per domain, but only for domains implemented and exported by those adapters. For example, `memory: new MemoryLibSQL(...)` and `workflows: new WorkflowsPG(...)` is valid because both packages export those domain classes.
 
 ## Usage
 
@@ -124,7 +128,9 @@ export const mastra = new Mastra({
 
 **default** (`MastraCompositeStore`): Default storage adapter. Domains not explicitly specified in \`domains\` will use this storage's domains as fallbacks.
 
-**domains** (`object`): Individual domain overrides. Each domain can come from a different storage adapter. These take precedence over the default storage.
+**disableInit** (`boolean`): When true, automatic initialization is disabled. You must call init() explicitly.
+
+**domains** (`object`): Individual domain overrides. Each domain can come from a different storage adapter. These take precedence over both \`editor\` and \`default\` storage.
 
 **domains.memory** (`MemoryStorage`): Storage for threads, messages, and resources.
 
@@ -136,7 +142,9 @@ export const mastra = new Mastra({
 
 **domains.agents** (`AgentsStorage`): Storage for stored agent configurations.
 
-**disableInit** (`boolean`): When true, automatic initialization is disabled. You must call init() explicitly.
+**domains.datasets** (`DatasetsStorage`): Storage for dataset metadata, dataset items, and dataset versions.
+
+**domains.experiments** (`ExperimentsStorage`): Storage for experiment runs and per-item experiment results.
 
 ## Initialization
 

package/dist/docs/references/reference-storage-dynamodb.md

@@ -1,4 +1,4 @@
-# DynamoDB Storage
+# DynamoDB storage
 
 The DynamoDB storage implementation provides a scalable and performant NoSQL database solution for Mastra, leveraging a single-table design pattern with [ElectroDB](https://electrodb.dev/).
 
@@ -120,7 +120,7 @@ For local development, you can use [DynamoDB Local](https://docs.aws.amazon.com/
 
 **config.ttl** (`object`): TTL (Time To Live) configuration for automatic data expiration. Configure per entity type: thread, message, trace, eval, workflow\_snapshot, resource, score. Each entity config includes: enabled (boolean), attributeName (string, default: 'ttl'), defaultTtlSeconds (number).
 
-## TTL (Time To Live) Configuration
+## TTL (time to live) configuration
 
 DynamoDB TTL allows you to automatically delete items after a specified time period. This is useful for:
 
@@ -216,7 +216,7 @@ aws dynamodb update-time-to-live \
 
 > **Note:** DynamoDB deletes expired items within 48 hours after expiration. Items remain queryable until actually deleted.
 
-## AWS IAM Permissions
+## AWS IAM permissions
 
 The IAM role or user executing the code needs appropriate permissions to interact with the specified DynamoDB table and its indexes. Below is a sample policy. Replace `${YOUR_TABLE_NAME}` with your actual table name and `${YOUR_AWS_REGION}` and `${YOUR_AWS_ACCOUNT_ID}` with appropriate values.
 
@@ -246,7 +246,7 @@ The IAM role or user executing the code needs appropriate permissions to interac
 }
 ```
 
-## Key Considerations
+## Key considerations
 
 Before diving into the architectural details, keep these key points in mind when working with the DynamoDB storage adapter:
 
@@ -255,7 +255,7 @@ Before diving into the architectural details, keep these key points in mind when
 - **Understanding GSIs:** Familiarity with how the GSIs are structured (as per `TABLE_SETUP.md`) is important for understanding data retrieval and potential query patterns.
 - **ElectroDB:** The adapter uses ElectroDB to manage interactions with DynamoDB, providing a layer of abstraction and type safety over raw DynamoDB operations.
 
-## Architectural Approach
+## Architectural approach
 
 This storage adapter utilizes a **single-table design pattern** leveraging [ElectroDB](https://electrodb.dev/), a common and recommended approach for DynamoDB. This differs architecturally from relational database adapters (like `@mastra/pg` or `@mastra/libsql`) that typically use multiple tables, each dedicated to a specific entity (threads, messages, etc.).
 

package/dist/docs/references/reference-storage-libsql.md

@@ -1,4 +1,4 @@
-# libSQL Storage
+# libSQL storage
 
 [libSQL](https://docs.turso.tech/libsql) is an open-source, SQLite-compatible database that supports both local and remote deployments. It can be used to store message history, workflow snapshots, traces, and eval scores.
 

package/dist/docs/references/reference-vectors-libsql.md

@@ -1,4 +1,4 @@
-# libSQL Vector Store
+# libSQL vector store
 
 The libSQL storage implementation provides a SQLite-compatible vector search [libSQL](https://github.com/tursodatabase/libsql), a fork of SQLite with vector extensions, and [Turso](https://turso.tech/) with vector extensions, offering a lightweight and efficient vector database solution. It's part of the `@mastra/libsql` package and offers efficient vector similarity search with metadata filtering.
 
@@ -69,7 +69,7 @@ const results = await store.query({
 });
 ```
 
-## Constructor Options
+## Constructor options
 
 **url** (`string`): libSQL database URL. Use ':memory:' for in-memory database, 'file:dbname.db' for local file, or a libSQL-compatible connection string like 'libsql://your-database.turso.io'.
 
@@ -81,7 +81,7 @@ const results = await store.query({
 
 ## Methods
 
-### createIndex()
+### `createIndex()`
 
 Creates a new vector collection. The index name must start with a letter or underscore and can only contain letters, numbers, and underscores. The dimension must be a positive integer.
 
@@ -91,7 +91,7 @@ Creates a new vector collection. The index name must start with a letter or unde
 
 **metric** (`'cosine' | 'euclidean' | 'dotproduct'`): Distance metric for similarity search. Note: Currently only cosine similarity is supported by libSQL. (Default: `cosine`)
 
-### upsert()
+### `upsert()`
 
 Adds or updates vectors and their metadata in the index. Uses a transaction to ensure all vectors are inserted atomically - if any insert fails, the entire operation is rolled back.
 
@@ -103,7 +103,7 @@ Adds or updates vectors and their metadata in the index. Uses a transaction to e
 
 **ids** (`string[]`): Optional vector IDs (auto-generated if not provided)
 
-### query()
+### `query()`
 
 Searches for similar vectors with optional metadata filtering.
 
@@ -119,7 +119,7 @@ Searches for similar vectors with optional metadata filtering.
 
 **minScore** (`number`): Minimum similarity score threshold (Default: `0`)
 
-### describeIndex()
+### `describeIndex()`
 
 Gets information about an index.
 
@@ -135,25 +135,25 @@ interface IndexStats {
 }
 ```
 
-### deleteIndex()
+### `deleteIndex()`
 
 Deletes an index and all its data.
 
 **indexName** (`string`): Name of the index to delete
 
-### listIndexes()
+### `listIndexes()`
 
 Lists all vector indexes in the database.
 
 Returns: `Promise<string[]>`
 
-### truncateIndex()
+### `truncateIndex()`
 
 Removes all vectors from an index while keeping the index structure.
 
 **indexName** (`string`): Name of the index to truncate
 
-### updateVector()
+### `updateVector()`
 
 Update a single vector by ID or by metadata filter. Either `id` or `filter` must be provided, but not both.
 
@@ -169,7 +169,7 @@ Update a single vector by ID or by metadata filter. Either `id` or `filter` must
 
 **update.metadata** (`Record<string, any>`): New metadata to update
 
-### deleteVector()
+### `deleteVector()`
 
 Deletes a specific vector entry from an index by its ID.
 
@@ -177,7 +177,7 @@ Deletes a specific vector entry from an index by its ID.
 
 **id** (`string`): ID of the vector entry to delete
 
-### deleteVectors()
+### `deleteVectors()`
 
 Delete multiple vectors by IDs or by metadata filter. Either `ids` or `filter` must be provided, but not both.
 
@@ -187,7 +187,7 @@ Delete multiple vectors by IDs or by metadata filter. Either `ids` or `filter` m
 
 **filter** (`Record<string, any>`): Metadata filter to identify vectors to delete (mutually exclusive with ids)
 
-## Response Types
+## Response types
 
 Query results are returned in this format:
 
@@ -200,7 +200,7 @@ interface QueryResult {
 }
 ```
 
-## Error Handling
+## Error handling
 
 The store throws specific errors for different failure cases:
 
@@ -232,7 +232,7 @@ Common error cases include:
 - Database connection issues
 - Transaction failures during upsert
 
-## Usage Example
+## Usage example
 
 ### Local embeddings with fastembed
 
@@ -277,7 +277,7 @@ export const libsqlAgent = new Agent({
   name: 'libSQL Agent',
   instructions:
     'You are an AI agent with the ability to automatically recall memories from previous interactions.',
-  model: 'openai/gpt-5.1',
+  model: 'openai/gpt-5.4',
   memory: new Memory({
     storage: new LibSQLStore({
       id: 'libsql-agent-storage',

package/dist/index.cjs

@@ -6994,15 +6994,15 @@ var MemoryLibSQL = class extends storage.MemoryStorage {
       let bestUnderTokens = 0;
       for (let i = 0; i < chunks.length; i++) {
         cumulativeMessageTokens += chunks[i].messageTokens ?? 0;
-        const boundary = i + 1;
+        const boundary2 = i + 1;
         if (cumulativeMessageTokens >= targetMessageTokens) {
           if (bestOverBoundary === 0 || cumulativeMessageTokens < bestOverTokens) {
-            bestOverBoundary = boundary;
+            bestOverBoundary = boundary2;
             bestOverTokens = cumulativeMessageTokens;
           }
         } else {
           if (cumulativeMessageTokens > bestUnderTokens) {
-            bestUnderBoundary = boundary;
+            bestUnderBoundary = boundary2;
             bestUnderTokens = cumulativeMessageTokens;
           }
         }
@@ -7037,9 +7037,12 @@ var MemoryLibSQL = class extends storage.MemoryStorage {
       const lastObservedAtStr = lastObservedAt.toISOString();
       const existingActive = row.activeObservations || "";
       const existingTokenCount = Number(row.observationTokenCount || 0);
-      const newActive = existingActive ? `${existingActive}
+      const boundary = `
 
-${activatedContent}` : activatedContent;
+--- message boundary (${lastObservedAt.toISOString()}) ---
+
+`;
+      const newActive = existingActive ? `${existingActive}${boundary}${activatedContent}` : activatedContent;
       const newTokenCount = existingTokenCount + activatedTokens;
       const existingPending = Number(row.pendingMessageTokens || 0);
       const newPending = Math.max(0, existingPending - activatedMessageTokens);