Kujo AI SDK 1.2.0 adds shared state

// 1h agoPRODUCT UPDATE

Kujo AI SDK 1.2.0 adds shared state

Kujo AI SDK v1.2.0 introduces process-local counters shared across client copies, plus inspection and reset APIs. It also improves usage accounting for completed model calls interrupted by budgets or structured-output handling.

// ANALYSIS

This is a practical reliability update that makes provider-agnostic AI integrations easier to observe and control in production.

–Shared counters reduce inconsistent usage tracking across copied clients
–Inspection and reset APIs simplify debugging, testing, and lifecycle management
–More accurate completed-call accounting improves budget and cost visibility
–Process-local state is useful for one runtime but is not a substitute for durable, cross-process billing telemetry

// TAGS

kujo-ai-sdksdkapiobservabilitystructured-outputagent-memory

DISCOVERED

1h ago

2026-10-13

PUBLISHED

1h ago

2026-10-13

RELEVANCE

8/ 10

AUTHOR

kujolang

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

NEWS24m ago

Open Together Rallies Open-Source AI Builders

Hugging Face’s Open Together brings AI builders to San Francisco on October 16 for community demos, open-model networking, and a dance party kicking off Open Source AI Week. Co-hosts span models, fine-tuning, inference, agents, code, and hardware.

LAUNCH42m ago

Merge Launches Workforce, Products AI Stack

Merge for Products gives teams authenticated agent tools, LLM routing, and customer-data sync, while Merge for Workforce governs employees’ AI access, models, MCP connectors, spending, and sensitive data. Together, they position Merge as infrastructure for both internal AI and customer-facing AI products. [Merge](https://www.merge.dev/)

BENCHMARK44m ago

ToneBench launches AI writing leaderboard

Towards AI has made ToneBench public, a benchmark measuring how well AI models write complete YouTube scripts in Louis-François Bouchard’s voice. Its leaderboard evaluates 78 models across 127 configurations using real production briefs, blind judging, and cost-per-script comparisons. [ToneBench](https://benchmark.towardsai.com/)