Repository navigation
ci: Version Packages - #2082
Merged
Merged
ci: Version Packages#2082
Conversation
Contributor
|
Important Review skippedBot user detected. To trigger a single review, invoke the
 |
github-actions
Bot
force-pushed
the
changeset-release/main
branch
5 times, most recently
from
November 9, 2026 03:25
9b960f4 to
567c06d
Compare
github-actions
Bot
force-pushed
the
changeset-release/main
branch
from
November 9, 2026 04:03
567c06d to
4b118a3
Compare
AlemTuzlak
approved these changes
Nov 9, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.
Releases
@tanstack/ai@1.3.0
Minor Changes
#2071
e8546f9- A run that the agent loop stops before the model finished now says so. WhenagentLoopStrategy(for examplemaxIterations) or a middlewareonShouldContinuereturnsfalsewhile tool results wait for the next model call,chat()ends the stream with aRUN_FINISHEDthat hasmetadata.tanstack.loopStopped: true, and the last stored assistant message gets the same field.finishReasonstays the reason of the last model call (tool_calls, orlengthfor a cut answer). A run that the model ends has noloopStopped.In
@tanstack/ai-harness, a turn that the loop stops settles with{ outcome: 'completed', loopStopped: true }.session.settled(), theharness.input.settledevent ofcreateHarnessClient().events(), and the transcript all carry it.#2087
5fd5355- Keep Responses item ids and phases when adjacent answer text is merged. Save each item's text length so replay can restore its boundary.Patch Changes
#2056
9393839-@tanstack/aiexportsstableJson: JSON text with the keys of every object in sorted order, for values that come back from a store with their keys reordered.@tanstack/ai-harnessnow imports it and no longer keeps its own copy.#2086
215e339- Stop a long-running server tool from leaking memory. While a tool ran, the event poll subscribed to the tool's promise again every 10 ms and kept every subscription until the tool finished. Memory grew by about 100 MB per hour while a tool stayed pending. The poll now subscribes once.#2069
7157ef3- Preserve metadata from live AG-UI TOOL_CALL_RESULT events on ToolResultPart.@tanstack/ai-harness@0.3.0
Minor Changes
#2071
e8546f9- A run that the agent loop stops before the model finished now says so. WhenagentLoopStrategy(for examplemaxIterations) or a middlewareonShouldContinuereturnsfalsewhile tool results wait for the next model call,chat()ends the stream with aRUN_FINISHEDthat hasmetadata.tanstack.loopStopped: true, and the last stored assistant message gets the same field.finishReasonstays the reason of the last model call (tool_calls, orlengthfor a cut answer). A run that the model ends has noloopStopped.In
@tanstack/ai-harness, a turn that the loop stops settles with{ outcome: 'completed', loopStopped: true }.session.settled(), theharness.input.settledevent ofcreateHarnessClient().events(), and the transcript all carry it.Patch Changes
#2056
9393839-@tanstack/aiexportsstableJson: JSON text with the keys of every object in sorted order, for values that come back from a store with their keys reordered.@tanstack/ai-harnessnow imports it and no longer keeps its own copy.Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-acp@1.0.2
Patch Changes
e8546f9,5fd5355,2753972,9393839,215e339,7157ef3]:@tanstack/ai-bedrock@1.2.1
Patch Changes
#2057
844496a- A Bedrock Converse chat stream that ends withstopReason: 'guardrail_intervened'now reportsfinishReason: 'content_filter'onRUN_FINISHED. Before, it was a normalstop. The structured-output stream uses the same finish reason when the body is valid JSON. When a guardrail leaves no JSON, that stream reportsRUN_ERRORwith codecontent_filterinstead ofempty-responseorparse-error.Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-code-mode@1.1.1
Patch Changes
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-compaction@1.0.1
Patch Changes
#2056
9393839- A compaction checkpoint is now reused when the message store gives messages back with their object keys in another order, as PostgreSQLjsonband schema parsing do.hashMessageshashes the messages withstableJson, so an equal history no longer looks changed and the summarizer does not run again over it. After the upgrade, saved checkpoints, saved usage counts and ready background summaries whose messages were not stored in sorted key order miss until the next compaction replaces them. Each such thread pays for one summarizer run.Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-computer-use@1.1.1
Patch Changes
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-gemini@1.2.1
Patch Changes
#2083
548ec42- Gemini Live (realtime)usageevents now includethoughtsTokenCountincompletionTokens, matching how the text adapters report usage.#2053
925a505- Send a multimodal tool result's parts to Gemini once, inside itsfunctionResponse. Before, the adapter also put them beside thefunctionResponsein the same turn, so each image was sent and billed twice.#2088
9bcdd41- Report Gemini content filter stops (SAFETY, RECITATION, BLOCKLIST, PROHIBITED_CONTENT, SPII, IMAGE_SAFETY, IMAGE_PROHIBITED_CONTENT, IMAGE_RECITATION) asfinishReason: "content_filter"on chat streams, and as acontent_filtererror fromstructuredOutputandstructuredOutputStream.Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-harness-cli@0.1.2
Patch Changes
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-mcp@1.1.1
Patch Changes
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-ollama@1.1.1
Patch Changes
#2077
c32f641- ReportfinishReason: "length"onRUN_FINISHEDwhen an Ollama stream ends withdone_reason: "length". This includes a reply cut bynum_predict, and a tool call cut the same way. A normal stop stays"stop". A finished tool call stays"tool_calls".Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-sandbox@1.2.1
Patch Changes
#2060
2753972- Keep the harness tool calls of earlier runs in a saved thread. Before,withSandboxremoved them from the chat transcript instead of only from the model request, so withwithPersistenceeach run's save dropped the tool calls of the runs before it, and a reload showed those runs without their tool cards.Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/ai-skills@1.0.2
Patch Changes
e8546f9,5fd5355,9393839,215e339,7157ef3]:@tanstack/openai-base@1.0.3
Patch Changes
#2087
5fd5355- Keep Responses item ids and phases when adjacent answer text is merged. Save each item's text length so replay can restore its boundary.Updated dependencies [
e8546f9,5fd5355,9393839,215e339,7157ef3]: