@cloudflare/ai-chat@0.8.3

cloudflare/agents@cloudflare/ai-chat@0.8.3Jun 5, 2026by github-actions[bot]

AI Summary

This patch release (@cloudflare/ai-chat@0.8.3) optimizes SQLite write performance by batching and packing chat-persistence operations, reducing database rows written and round-trips by up to ~10×. The changes include packing stream chunks into single SQLite rows, batched DELETE queries for message cleanup and submission deletion, and batched incident TTL sweeps.

Key Highlights

ResumableStream now packs buffered stream chunks into single SQLite rows (JSON arrays) instead of one row per chunk, reducing storage by ~10×
Message cleanup now uses batched DELETE ... WHERE id IN (...) queries instead of individual DELETE statements per row
@cloudflare/think deleteSubmissions() now batch deletes terminal submissions in a single query
Chat-recovery incident TTL sweep now uses batched storage.delete(keys) calls (up to 128 keys) instead of individual awaited deletes
Backward compatible: reads transparently unpack both new packed segments and legacy per-chunk rows

New Features

Packed SQLite storage for stream chunks (JSON arrays of chunk bodies) with per-segment byte cap for 2 MB limit
Batched DELETE queries with configurable bound parameter limits (100 for messages, full batch for submissions)
Batched storage.delete() calls for incident cleanup (up to 128 keys per call)
New shared helpers exported: buildInClauseStrings and MAX_BOUND_PARAMS from agents/chat

Full Release Notes

### Patch Changes

-   [#1686](https://github.com/cloudflare/agents/pull/1686) [`1e49880`](https://github.com/cloudflare/agents/commit/1e498803fe26970aa264678d5ae3a2c96dd28258) Thanks [@threepointone](https://github.com/threepointone)! - Batch and pack chat-persistence SQLite writes to reduce rows written and round-trips.
    -   `agents`: `ResumableStream` now **packs** each buffered group of stream chunks into a single SQLite row (a JSON array of chunk bodies) instead of writing one row per chunk. Single-chunk and large-chunk segments are stored unwrapped, and a per-segment byte cap keeps rows within the 2 MB SQLite row limit. This cuts chunk rows written / stored / scanned-on-replay by up to ~10×. Reads (replay, orphan reconstruction, `getStreamChunks`) transparently unpack both packed segments and legacy per-chunk rows, so existing stored data keeps working. Adds shared `buildInClauseStrings` and `MAX_BOUND_PARAMS` helpers exported from `agents/chat`.
    -   `@cloudflare/ai-chat`: message cleanup (stale-row pruning and `maxPersistedMessages` enforcement) previously issued one `DELETE` per row in a loop; it now deletes rows in batched `DELETE ... WHERE id IN (...)` queries (capped at 100 bound parameters per query).
    -   `@cloudflare/think`: `deleteSubmissions()` cleanup previously issued one `DELETE` per terminal submission (up to 500 per call); it now deletes rows in batched `DELETE ... WHERE submission_id IN (...)` queries.
    -   `@cloudflare/ai-chat` & `@cloudflare/think`: chat-recovery incident TTL sweep previously deleted each stale incident with a separate awaited `storage.delete(key)` (which also defeats Durable Object write-coalescing); it now deletes incidents in batched `storage.delete(keys)` calls (up to 128 keys per call).