TanStackOSS
    ai
    feat(ai): add shared Scope identity type Introduce a single identity/isolation vocabulary (Scope) in @tanstack/ai for the subsystems that persist or recall per-conversation data — ai-persistence and ai-memory. threadId is the one conversation key (matching ChatMiddlewareContext.threadId); userId/tenantId/namespace are optional. Every field is an isolation boundary derived server-side, never from client input. Additive precursor: nothing consumes Scope yet. Landed ahead of the persistence (#785) and memory (#541) PRs so both build on one settled identity contract instead of diverging (sessionId vs threadId, bare threadId vs flat scope:string).
    nx run-many --targets=build --exclude=examples/**,testing/**
Succeeded
CI Pipeline Execution
Linux
4 CPU cores
read-write access token used
220ce141980
4%

Cache hits

2 of 48 tasks used cache.

Get faster results

Learn how to enable distribution.

Atomizer enabled

48 groups of tasks optimized.

© 2026 - Nx Cloud

Terms of ServicePrivacy PolicyChangelogStatusDocsContact Nx CloudPricingCompany@NxDevTools